28 Commits
v9.2 ... v11.3

Author SHA1 Message Date
Rutwik Choughule
568b39bd5b add and update samples with CUDA 11.3 support 2021-04-16 11:54:26 +05:30
Mahesh Doijade
067cb65523 -- Add partitioned cuda pipeline prod-cons gemm kernel
-- Add cudaCompressibleMemory sample to use copy engine vs SM writes
   depending on arch
2021-03-03 22:53:02 +05:30
Mahesh Doijade
b882fa00ee -- Add freeglut and glew64 libs for simpleGL
-- delete freeimage libs and headers, user needs to install it for
building samples depending on it
2020-12-11 17:34:13 +05:30
Mahesh Doijade
1cd3264681 Add and update samples with CUDA 11.2 support 2020-12-10 01:05:32 +05:30
Mahesh Doijade
92b0568792 Add cudaNvSciNvMedia sample with/without nvsci* APIs, it takes RGBA image as input and produces YUV via nvmedia
this YUV is consumed by cuda which converts it to grayscale image which is written to file as output
2020-11-24 16:28:04 +05:30
Mahesh Doijade
c4e2869a2b add multi-warp cooperative groups based reduction kernel in reduction sample 2020-09-24 16:49:58 +05:30
Mahesh Doijade
dd2dba3489 helper_string: fix the missing "/" at end of 3 up path in sdkFindFilePath() 2020-09-17 20:45:23 +05:30
Mahesh Doijade
44052982a7 add ewp tags to VS project of conjugateGradientMultiDeviceCG 2020-09-17 20:01:41 +05:30
Mahesh Doijade
cd76533c3f Add and update samples for cuda 11.1 support 2020-09-15 23:45:56 +05:30
Mahesh Doijade
e6ce58fef4 [p2pBandwidthLatency] increase default buffer size and add support to specific buffer size through command line 2020-06-30 19:05:55 +05:30
Mahesh Doijade
0ec4bd58e5 cudaCompressibleMemory: when unable to allocate compressible memory waive the execution 2020-06-01 20:22:22 +05:30
Mahesh Doijade
908dddb207 cudaCompressibleMemory: refactor and refine, and use only cuMemMap API for non-compressible allocation.
hence, remove use of cudaMalloc.
2020-05-27 19:00:33 +05:30
Mahesh Doijade
b60c149b17 Correct the runtime API section in documentation when sample has both driver as well as runtime API 2020-05-22 22:33:02 +05:30
Mahesh Doijade
56b5b4c767 -- add cudaCompressibleMemory sample
-- fix simpleD3D12 build issue
-- fix simpleVulkanMMAP shader file locating issue
2020-05-22 22:07:30 +05:30
Mahesh Doijade
53e9e413a8 Update VS project files to use cuda 11.0 props/targets 2020-05-19 18:16:05 +05:30
Mahesh Doijade
4f6e02970b Add and update samples for cuda 11.0 support 2020-05-18 22:22:06 +05:30
Mahesh Doijade
6be514679b Add and update samples for cuda 10.2 support 2019-10-23 19:00:39 +05:30
Mahesh Doijade
489d9f7b1f conjugateGradientCudaGraphs: Add cusparseScsrmv & cublasSdot instead of
gpuSpMV & gpuDotProduct as they are graph compliant
2019-08-19 15:47:24 +05:30
Mahesh Doijade
c8483e0798 Add release notes for cuda 10.1 update 2 support 2019-08-05 18:58:54 +05:30
Mahesh Doijade
58d0ceff18 Add and update samples for cuda 10.1 update 2 support 2019-08-02 23:15:47 +05:30
Mahesh Doijade
337815dbee Add and update samples with CUDA 10.1 Update 1 support 2019-04-10 20:12:09 +05:30
Mahesh Doijade
1abc294982 --Update README.md for cuda 10.1 release
-- fix multidevice CG sample to check for concurrentManagedAccess as it needs systemwide atomics
2019-01-23 16:56:20 +05:30
Mahesh Doijade
b458dafcd6 Add and update samples with CUDA 10.1 support 2019-01-23 01:34:43 +05:30
Mahesh Doijade
32f0fc6111 fix the description of CUDA IPC as it is supported on Windows as well since CUDA 9.2 2018-10-24 13:35:39 +05:30
Mahesh Doijade
6bbaf9ac80 -- use DX_SDK env variable only for non-DX12 samples VS files
as DX12 does not need legacy DX SDK
2018-10-09 22:42:25 +05:30
Mahesh Doijade
fcb23487a8 -- Add simpleD3D12 sample which demonstrates DX12-CUDA interop
rendering sinewave
-- Update the documentation with description about this sample
2018-10-04 20:05:23 +05:30
Mahesh Doijade
91dc60dd96 simpleVulkan: windows.h should be included before VersionHelpers.h 2018-09-10 20:41:05 +05:30
Mahesh Doijade
21c36d3568 Add and Update samples for CUDA 10.0 2018-08-24 22:35:15 +05:30
679 changed files with 217080 additions and 4963 deletions

22
Common/GL/freeglut.h Normal file
View File

@@ -0,0 +1,22 @@
#ifndef __FREEGLUT_H__
#define __FREEGLUT_H__
/*
* freeglut.h
*
* The freeglut library include file
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* PAWEL W. OLSZTA BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
* IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#include "freeglut_std.h"
#include "freeglut_ext.h"
/*** END OF FILE ***/
#endif /* __FREEGLUT_H__ */

115
Common/GL/freeglut_ext.h Normal file
View File

@@ -0,0 +1,115 @@
#ifndef __FREEGLUT_EXT_H__
#define __FREEGLUT_EXT_H__
/*
* freeglut_ext.h
*
* The non-GLUT-compatible extensions to the freeglut library include file
*
* Copyright (c) 1999-2000 Pawel W. Olszta. All Rights Reserved.
* Written by Pawel W. Olszta, <olszta@sourceforge.net>
* Creation date: Thu Dec 2 1999
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included
* in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* PAWEL W. OLSZTA BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
* IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#ifdef __cplusplus
extern "C" {
#endif
/*
* GLUT API Extension macro definitions -- behaviour when the user clicks on an "x" to close a window
*/
#define GLUT_ACTION_EXIT 0
#define GLUT_ACTION_GLUTMAINLOOP_RETURNS 1
#define GLUT_ACTION_CONTINUE_EXECUTION 2
/*
* Create a new rendering context when the user opens a new window?
*/
#define GLUT_CREATE_NEW_CONTEXT 0
#define GLUT_USE_CURRENT_CONTEXT 1
/*
* GLUT API Extension macro definitions -- the glutGet parameters
*/
#define GLUT_ACTION_ON_WINDOW_CLOSE 0x01F9
#define GLUT_WINDOW_BORDER_WIDTH 0x01FA
#define GLUT_WINDOW_HEADER_HEIGHT 0x01FB
#define GLUT_VERSION 0x01FC
#define GLUT_RENDERING_CONTEXT 0x01FD
/*
* Process loop function, see freeglut_main.c
*/
FGAPI void FGAPIENTRY glutMainLoopEvent(void);
FGAPI void FGAPIENTRY glutLeaveMainLoop(void);
/*
* Window-specific callback functions, see freeglut_callbacks.c
*/
FGAPI void FGAPIENTRY glutMouseWheelFunc(void (* callback)(int, int, int, int));
FGAPI void FGAPIENTRY glutCloseFunc(void (* callback)(void));
FGAPI void FGAPIENTRY glutWMCloseFunc(void (* callback)(void));
/* A. Donev: Also a destruction callback for menus */
FGAPI void FGAPIENTRY glutMenuDestroyFunc(void (* callback)(void));
/*
* State setting and retrieval functions, see freeglut_state.c
*/
FGAPI void FGAPIENTRY glutSetOption(GLenum option_flag, int value) ;
/* A.Donev: User-data manipulation */
FGAPI void *FGAPIENTRY glutGetWindowData(void);
FGAPI void FGAPIENTRY glutSetWindowData(void *data);
FGAPI void *FGAPIENTRY glutGetMenuData(void);
FGAPI void FGAPIENTRY glutSetMenuData(void *data);
/*
* Font stuff, see freeglut_font.c
*/
FGAPI int FGAPIENTRY glutBitmapHeight(void *font);
FGAPI GLfloat FGAPIENTRY glutStrokeHeight(void *font);
FGAPI void FGAPIENTRY glutBitmapString(void *font, const unsigned char *string);
FGAPI void FGAPIENTRY glutStrokeString(void *font, const unsigned char *string);
/*
* Geometry functions, see freeglut_geometry.c
*/
FGAPI void FGAPIENTRY glutWireRhombicDodecahedron(void);
FGAPI void FGAPIENTRY glutSolidRhombicDodecahedron(void);
FGAPI void FGAPIENTRY glutWireSierpinskiSponge(int num_levels, GLdouble offset[3], GLdouble scale) ;
FGAPI void FGAPIENTRY glutSolidSierpinskiSponge(int num_levels, GLdouble offset[3], GLdouble scale) ;
FGAPI void FGAPIENTRY glutWireCylinder(GLdouble radius, GLdouble height, GLint slices, GLint stacks);
FGAPI void FGAPIENTRY glutSolidCylinder(GLdouble radius, GLdouble height, GLint slices, GLint stacks);
/*
* Extension functions, see freeglut_ext.c
*/
FGAPI void *FGAPIENTRY glutGetProcAddress(const char *procName);
#ifdef __cplusplus
}
#endif
/*** END OF FILE ***/
#endif /* __FREEGLUT_EXT_H__ */

547
Common/GL/freeglut_std.h Normal file
View File

@@ -0,0 +1,547 @@
#ifndef __FREEGLUT_STD_H__
#define __FREEGLUT_STD_H__
/*
* freeglut_std.h
*
* The GLUT-compatible part of the freeglut library include file
*
* Copyright (c) 1999-2000 Pawel W. Olszta. All Rights Reserved.
* Written by Pawel W. Olszta, <olszta@sourceforge.net>
* Creation date: Thu Dec 2 1999
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included
* in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* PAWEL W. OLSZTA BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
* IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#ifdef __cplusplus
extern "C" {
#endif
/*
* Under windows, we have to differentiate between static and dynamic libraries
*/
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
# include <windows.h>
# include <windowsx.h>
# include <mmsystem.h>
# define WINDOWS
#ifdef FREEGLUT_STATIC
# define FGAPI
# define FGAPIENTRY
# pragma comment (lib, "freeglut_static.lib") /* link with Win32 static freeglut lib */
#else
# if defined(FREEGLUT_EXPORTS)
# define FGAPI __declspec(dllexport)
/* # define FGAPI */
# else
# define FGAPI __declspec(dllimport)
# pragma comment (lib, "freeglut.lib") /* link with Win32 freeglut lib */
# endif
# define FGAPIENTRY __stdcall
#endif
#pragma comment (lib, "winmm.lib") /* link with Windows MultiMedia lib */
#pragma comment (lib, "user32.lib") /* link with Windows user lib */
#pragma comment (lib, "gdi32.lib") /* link with Windows GDI lib */
#pragma comment (lib, "opengl32.lib") /* link with Microsoft OpenGL lib */
#pragma comment (lib, "glu32.lib") /* link with OpenGL Utility lib */
#else
# define FGAPI
# define FGAPIENTRY
#endif
/*
* The freeglut and GLUT API versions
*/
#define FREEGLUT 1
#define GLUT_API_VERSION 4
#define FREEGLUT_VERSION_2_0 1
/*
* Always include OpenGL and GLU headers
*/
#include <GL/gl.h>
#include <GL/glu.h>
/*
* GLUT API macro definitions -- the special key codes:
*/
#define GLUT_KEY_F1 0x0001
#define GLUT_KEY_F2 0x0002
#define GLUT_KEY_F3 0x0003
#define GLUT_KEY_F4 0x0004
#define GLUT_KEY_F5 0x0005
#define GLUT_KEY_F6 0x0006
#define GLUT_KEY_F7 0x0007
#define GLUT_KEY_F8 0x0008
#define GLUT_KEY_F9 0x0009
#define GLUT_KEY_F10 0x000A
#define GLUT_KEY_F11 0x000B
#define GLUT_KEY_F12 0x000C
#define GLUT_KEY_LEFT 0x0064
#define GLUT_KEY_UP 0x0065
#define GLUT_KEY_RIGHT 0x0066
#define GLUT_KEY_DOWN 0x0067
#define GLUT_KEY_PAGE_UP 0x0068
#define GLUT_KEY_PAGE_DOWN 0x0069
#define GLUT_KEY_HOME 0x006A
#define GLUT_KEY_END 0x006B
#define GLUT_KEY_INSERT 0x006C
/*
* GLUT API macro definitions -- mouse state definitions
*/
#define GLUT_LEFT_BUTTON 0x0000
#define GLUT_MIDDLE_BUTTON 0x0001
#define GLUT_RIGHT_BUTTON 0x0002
#define GLUT_DOWN 0x0000
#define GLUT_UP 0x0001
#define GLUT_LEFT 0x0000
#define GLUT_ENTERED 0x0001
/*
* GLUT API macro definitions -- the display mode definitions
*/
#define GLUT_RGB 0x0000
#define GLUT_RGBA 0x0000
#define GLUT_INDEX 0x0001
#define GLUT_SINGLE 0x0000
#define GLUT_DOUBLE 0x0002
#define GLUT_ACCUM 0x0004
#define GLUT_ALPHA 0x0008
#define GLUT_DEPTH 0x0010
#define GLUT_STENCIL 0x0020
#define GLUT_MULTISAMPLE 0x0080
#define GLUT_STEREO 0x0100
#define GLUT_LUMINANCE 0x0200
/*
* GLUT API macro definitions -- windows and menu related definitions
*/
#define GLUT_MENU_NOT_IN_USE 0x0000
#define GLUT_MENU_IN_USE 0x0001
#define GLUT_NOT_VISIBLE 0x0000
#define GLUT_VISIBLE 0x0001
#define GLUT_HIDDEN 0x0000
#define GLUT_FULLY_RETAINED 0x0001
#define GLUT_PARTIALLY_RETAINED 0x0002
#define GLUT_FULLY_COVERED 0x0003
/*
* GLUT API macro definitions -- fonts definitions
*
* Steve Baker suggested to make it binary compatible with GLUT:
*/
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
# define GLUT_STROKE_ROMAN ((void *)0x0000)
# define GLUT_STROKE_MONO_ROMAN ((void *)0x0001)
# define GLUT_BITMAP_9_BY_15 ((void *)0x0002)
# define GLUT_BITMAP_8_BY_13 ((void *)0x0003)
# define GLUT_BITMAP_TIMES_ROMAN_10 ((void *)0x0004)
# define GLUT_BITMAP_TIMES_ROMAN_24 ((void *)0x0005)
# define GLUT_BITMAP_HELVETICA_10 ((void *)0x0006)
# define GLUT_BITMAP_HELVETICA_12 ((void *)0x0007)
# define GLUT_BITMAP_HELVETICA_18 ((void *)0x0008)
#else
/*
* I don't really know if it's a good idea... But here it goes:
*/
extern void *glutStrokeRoman;
extern void *glutStrokeMonoRoman;
extern void *glutBitmap9By15;
extern void *glutBitmap8By13;
extern void *glutBitmapTimesRoman10;
extern void *glutBitmapTimesRoman24;
extern void *glutBitmapHelvetica10;
extern void *glutBitmapHelvetica12;
extern void *glutBitmapHelvetica18;
/*
* Those pointers will be used by following definitions:
*/
# define GLUT_STROKE_ROMAN ((void *) &glutStrokeRoman)
# define GLUT_STROKE_MONO_ROMAN ((void *) &glutStrokeMonoRoman)
# define GLUT_BITMAP_9_BY_15 ((void *) &glutBitmap9By15)
# define GLUT_BITMAP_8_BY_13 ((void *) &glutBitmap8By13)
# define GLUT_BITMAP_TIMES_ROMAN_10 ((void *) &glutBitmapTimesRoman10)
# define GLUT_BITMAP_TIMES_ROMAN_24 ((void *) &glutBitmapTimesRoman24)
# define GLUT_BITMAP_HELVETICA_10 ((void *) &glutBitmapHelvetica10)
# define GLUT_BITMAP_HELVETICA_12 ((void *) &glutBitmapHelvetica12)
# define GLUT_BITMAP_HELVETICA_18 ((void *) &glutBitmapHelvetica18)
#endif
/*
* GLUT API macro definitions -- the glutGet parameters
*/
#define GLUT_WINDOW_X 0x0064
#define GLUT_WINDOW_Y 0x0065
#define GLUT_WINDOW_WIDTH 0x0066
#define GLUT_WINDOW_HEIGHT 0x0067
#define GLUT_WINDOW_BUFFER_SIZE 0x0068
#define GLUT_WINDOW_STENCIL_SIZE 0x0069
#define GLUT_WINDOW_DEPTH_SIZE 0x006A
#define GLUT_WINDOW_RED_SIZE 0x006B
#define GLUT_WINDOW_GREEN_SIZE 0x006C
#define GLUT_WINDOW_BLUE_SIZE 0x006D
#define GLUT_WINDOW_ALPHA_SIZE 0x006E
#define GLUT_WINDOW_ACCUM_RED_SIZE 0x006F
#define GLUT_WINDOW_ACCUM_GREEN_SIZE 0x0070
#define GLUT_WINDOW_ACCUM_BLUE_SIZE 0x0071
#define GLUT_WINDOW_ACCUM_ALPHA_SIZE 0x0072
#define GLUT_WINDOW_DOUBLEBUFFER 0x0073
#define GLUT_WINDOW_RGBA 0x0074
#define GLUT_WINDOW_PARENT 0x0075
#define GLUT_WINDOW_NUM_CHILDREN 0x0076
#define GLUT_WINDOW_COLORMAP_SIZE 0x0077
#define GLUT_WINDOW_NUM_SAMPLES 0x0078
#define GLUT_WINDOW_STEREO 0x0079
#define GLUT_WINDOW_CURSOR 0x007A
#define GLUT_SCREEN_WIDTH 0x00C8
#define GLUT_SCREEN_HEIGHT 0x00C9
#define GLUT_SCREEN_WIDTH_MM 0x00CA
#define GLUT_SCREEN_HEIGHT_MM 0x00CB
#define GLUT_MENU_NUM_ITEMS 0x012C
#define GLUT_DISPLAY_MODE_POSSIBLE 0x0190
#define GLUT_INIT_WINDOW_X 0x01F4
#define GLUT_INIT_WINDOW_Y 0x01F5
#define GLUT_INIT_WINDOW_WIDTH 0x01F6
#define GLUT_INIT_WINDOW_HEIGHT 0x01F7
#define GLUT_INIT_DISPLAY_MODE 0x01F8
#define GLUT_ELAPSED_TIME 0x02BC
#define GLUT_WINDOW_FORMAT_ID 0x007B
#define GLUT_INIT_STATE 0x007C
/*
* GLUT API macro definitions -- the glutDeviceGet parameters
*/
#define GLUT_HAS_KEYBOARD 0x0258
#define GLUT_HAS_MOUSE 0x0259
#define GLUT_HAS_SPACEBALL 0x025A
#define GLUT_HAS_DIAL_AND_BUTTON_BOX 0x025B
#define GLUT_HAS_TABLET 0x025C
#define GLUT_NUM_MOUSE_BUTTONS 0x025D
#define GLUT_NUM_SPACEBALL_BUTTONS 0x025E
#define GLUT_NUM_BUTTON_BOX_BUTTONS 0x025F
#define GLUT_NUM_DIALS 0x0260
#define GLUT_NUM_TABLET_BUTTONS 0x0261
#define GLUT_DEVICE_IGNORE_KEY_REPEAT 0x0262
#define GLUT_DEVICE_KEY_REPEAT 0x0263
#define GLUT_HAS_JOYSTICK 0x0264
#define GLUT_OWNS_JOYSTICK 0x0265
#define GLUT_JOYSTICK_BUTTONS 0x0266
#define GLUT_JOYSTICK_AXES 0x0267
#define GLUT_JOYSTICK_POLL_RATE 0x0268
/*
* GLUT API macro definitions -- the glutLayerGet parameters
*/
#define GLUT_OVERLAY_POSSIBLE 0x0320
#define GLUT_LAYER_IN_USE 0x0321
#define GLUT_HAS_OVERLAY 0x0322
#define GLUT_TRANSPARENT_INDEX 0x0323
#define GLUT_NORMAL_DAMAGED 0x0324
#define GLUT_OVERLAY_DAMAGED 0x0325
/*
* GLUT API macro definitions -- the glutVideoResizeGet parameters
*/
#define GLUT_VIDEO_RESIZE_POSSIBLE 0x0384
#define GLUT_VIDEO_RESIZE_IN_USE 0x0385
#define GLUT_VIDEO_RESIZE_X_DELTA 0x0386
#define GLUT_VIDEO_RESIZE_Y_DELTA 0x0387
#define GLUT_VIDEO_RESIZE_WIDTH_DELTA 0x0388
#define GLUT_VIDEO_RESIZE_HEIGHT_DELTA 0x0389
#define GLUT_VIDEO_RESIZE_X 0x038A
#define GLUT_VIDEO_RESIZE_Y 0x038B
#define GLUT_VIDEO_RESIZE_WIDTH 0x038C
#define GLUT_VIDEO_RESIZE_HEIGHT 0x038D
/*
* GLUT API macro definitions -- the glutUseLayer parameters
*/
#define GLUT_NORMAL 0x0000
#define GLUT_OVERLAY 0x0001
/*
* GLUT API macro definitions -- the glutGetModifiers parameters
*/
#define GLUT_ACTIVE_SHIFT 0x0001
#define GLUT_ACTIVE_CTRL 0x0002
#define GLUT_ACTIVE_ALT 0x0004
/*
* GLUT API macro definitions -- the glutSetCursor parameters
*/
#define GLUT_CURSOR_RIGHT_ARROW 0x0000
#define GLUT_CURSOR_LEFT_ARROW 0x0001
#define GLUT_CURSOR_INFO 0x0002
#define GLUT_CURSOR_DESTROY 0x0003
#define GLUT_CURSOR_HELP 0x0004
#define GLUT_CURSOR_CYCLE 0x0005
#define GLUT_CURSOR_SPRAY 0x0006
#define GLUT_CURSOR_WAIT 0x0007
#define GLUT_CURSOR_TEXT 0x0008
#define GLUT_CURSOR_CROSSHAIR 0x0009
#define GLUT_CURSOR_UP_DOWN 0x000A
#define GLUT_CURSOR_LEFT_RIGHT 0x000B
#define GLUT_CURSOR_TOP_SIDE 0x000C
#define GLUT_CURSOR_BOTTOM_SIDE 0x000D
#define GLUT_CURSOR_LEFT_SIDE 0x000E
#define GLUT_CURSOR_RIGHT_SIDE 0x000F
#define GLUT_CURSOR_TOP_LEFT_CORNER 0x0010
#define GLUT_CURSOR_TOP_RIGHT_CORNER 0x0011
#define GLUT_CURSOR_BOTTOM_RIGHT_CORNER 0x0012
#define GLUT_CURSOR_BOTTOM_LEFT_CORNER 0x0013
#define GLUT_CURSOR_INHERIT 0x0064
#define GLUT_CURSOR_NONE 0x0065
#define GLUT_CURSOR_FULL_CROSSHAIR 0x0066
/*
* GLUT API macro definitions -- RGB color component specification definitions
*/
#define GLUT_RED 0x0000
#define GLUT_GREEN 0x0001
#define GLUT_BLUE 0x0002
/*
* GLUT API macro definitions -- additional keyboard and joystick definitions
*/
#define GLUT_KEY_REPEAT_OFF 0x0000
#define GLUT_KEY_REPEAT_ON 0x0001
#define GLUT_KEY_REPEAT_DEFAULT 0x0002
#define GLUT_JOYSTICK_BUTTON_A 0x0001
#define GLUT_JOYSTICK_BUTTON_B 0x0002
#define GLUT_JOYSTICK_BUTTON_C 0x0004
#define GLUT_JOYSTICK_BUTTON_D 0x0008
/*
* GLUT API macro definitions -- game mode definitions
*/
#define GLUT_GAME_MODE_ACTIVE 0x0000
#define GLUT_GAME_MODE_POSSIBLE 0x0001
#define GLUT_GAME_MODE_WIDTH 0x0002
#define GLUT_GAME_MODE_HEIGHT 0x0003
#define GLUT_GAME_MODE_PIXEL_DEPTH 0x0004
#define GLUT_GAME_MODE_REFRESH_RATE 0x0005
#define GLUT_GAME_MODE_DISPLAY_CHANGED 0x0006
/*
* Initialization functions, see fglut_init.c
*/
FGAPI void FGAPIENTRY glutInit(int *pargc, char **argv);
FGAPI void FGAPIENTRY glutInitWindowPosition(int x, int y);
FGAPI void FGAPIENTRY glutInitWindowSize(int width, int height);
FGAPI void FGAPIENTRY glutInitDisplayMode(unsigned int displayMode);
FGAPI void FGAPIENTRY glutInitDisplayString(const char *displayMode);
/*
* Process loop function, see freeglut_main.c
*/
FGAPI void FGAPIENTRY glutMainLoop(void);
/*
* Window management functions, see freeglut_window.c
*/
FGAPI int FGAPIENTRY glutCreateWindow(const char *title);
FGAPI int FGAPIENTRY glutCreateSubWindow(int window, int x, int y, int width, int height);
FGAPI void FGAPIENTRY glutDestroyWindow(int window);
FGAPI void FGAPIENTRY glutSetWindow(int window);
FGAPI int FGAPIENTRY glutGetWindow(void);
FGAPI void FGAPIENTRY glutSetWindowTitle(const char *title);
FGAPI void FGAPIENTRY glutSetIconTitle(const char *title);
FGAPI void FGAPIENTRY glutReshapeWindow(int width, int height);
FGAPI void FGAPIENTRY glutPositionWindow(int x, int y);
FGAPI void FGAPIENTRY glutShowWindow(void);
FGAPI void FGAPIENTRY glutHideWindow(void);
FGAPI void FGAPIENTRY glutIconifyWindow(void);
FGAPI void FGAPIENTRY glutPushWindow(void);
FGAPI void FGAPIENTRY glutPopWindow(void);
FGAPI void FGAPIENTRY glutFullScreen(void);
/*
* Display-connected functions, see freeglut_display.c
*/
FGAPI void FGAPIENTRY glutPostWindowRedisplay(int window);
FGAPI void FGAPIENTRY glutPostRedisplay(void);
FGAPI void FGAPIENTRY glutSwapBuffers(void);
/*
* Mouse cursor functions, see freeglut_cursor.c
*/
FGAPI void FGAPIENTRY glutWarpPointer(int x, int y);
FGAPI void FGAPIENTRY glutSetCursor(int cursor);
/*
* Overlay stuff, see freeglut_overlay.c
*/
FGAPI void FGAPIENTRY glutEstablishOverlay(void);
FGAPI void FGAPIENTRY glutRemoveOverlay(void);
FGAPI void FGAPIENTRY glutUseLayer(GLenum layer);
FGAPI void FGAPIENTRY glutPostOverlayRedisplay(void);
FGAPI void FGAPIENTRY glutPostWindowOverlayRedisplay(int window);
FGAPI void FGAPIENTRY glutShowOverlay(void);
FGAPI void FGAPIENTRY glutHideOverlay(void);
/*
* Menu stuff, see freeglut_menu.c
*/
FGAPI int FGAPIENTRY glutCreateMenu(void (* callback)(int menu));
FGAPI void FGAPIENTRY glutDestroyMenu(int menu);
FGAPI int FGAPIENTRY glutGetMenu(void);
FGAPI void FGAPIENTRY glutSetMenu(int menu);
FGAPI void FGAPIENTRY glutAddMenuEntry(const char *label, int value);
FGAPI void FGAPIENTRY glutAddSubMenu(const char *label, int subMenu);
FGAPI void FGAPIENTRY glutChangeToMenuEntry(int item, const char *label, int value);
FGAPI void FGAPIENTRY glutChangeToSubMenu(int item, const char *label, int value);
FGAPI void FGAPIENTRY glutRemoveMenuItem(int item);
FGAPI void FGAPIENTRY glutAttachMenu(int button);
FGAPI void FGAPIENTRY glutDetachMenu(int button);
/*
* Global callback functions, see freeglut_callbacks.c
*/
FGAPI void FGAPIENTRY glutTimerFunc(unsigned int time, void (* callback)(int), int value);
FGAPI void FGAPIENTRY glutIdleFunc(void (* callback)(void));
/*
* Window-specific callback functions, see freeglut_callbacks.c
*/
FGAPI void FGAPIENTRY glutKeyboardFunc(void (* callback)(unsigned char, int, int));
FGAPI void FGAPIENTRY glutSpecialFunc(void (* callback)(int, int, int));
FGAPI void FGAPIENTRY glutReshapeFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutVisibilityFunc(void (* callback)(int));
FGAPI void FGAPIENTRY glutDisplayFunc(void (* callback)(void));
FGAPI void FGAPIENTRY glutMouseFunc(void (* callback)(int, int, int, int));
FGAPI void FGAPIENTRY glutMotionFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutPassiveMotionFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutEntryFunc(void (* callback)(int));
FGAPI void FGAPIENTRY glutKeyboardUpFunc(void (* callback)(unsigned char, int, int));
FGAPI void FGAPIENTRY glutSpecialUpFunc(void (* callback)(int, int, int));
FGAPI void FGAPIENTRY glutJoystickFunc(void (* callback)(unsigned int, int, int, int), int pollInterval);
FGAPI void FGAPIENTRY glutMenuStateFunc(void (* callback)(int));
FGAPI void FGAPIENTRY glutMenuStatusFunc(void (* callback)(int, int, int));
FGAPI void FGAPIENTRY glutOverlayDisplayFunc(void (* callback)(void));
FGAPI void FGAPIENTRY glutWindowStatusFunc(void (* callback)(int));
FGAPI void FGAPIENTRY glutSpaceballMotionFunc(void (* callback)(int, int, int));
FGAPI void FGAPIENTRY glutSpaceballRotateFunc(void (* callback)(int, int, int));
FGAPI void FGAPIENTRY glutSpaceballButtonFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutButtonBoxFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutDialsFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutTabletMotionFunc(void (* callback)(int, int));
FGAPI void FGAPIENTRY glutTabletButtonFunc(void (* callback)(int, int, int, int));
/*
* State setting and retrieval functions, see freeglut_state.c
*/
FGAPI int FGAPIENTRY glutGet(GLenum query);
FGAPI int FGAPIENTRY glutDeviceGet(GLenum query);
FGAPI int FGAPIENTRY glutGetModifiers(void);
FGAPI int FGAPIENTRY glutLayerGet(GLenum query);
/*
* Font stuff, see freeglut_font.c
*/
FGAPI void FGAPIENTRY glutBitmapCharacter(void *font, int character);
FGAPI int FGAPIENTRY glutBitmapWidth(void *font, int character);
FGAPI void FGAPIENTRY glutStrokeCharacter(void *font, int character);
FGAPI int FGAPIENTRY glutStrokeWidth(void *font, int character);
FGAPI int FGAPIENTRY glutBitmapLength(void *font, const unsigned char *string);
FGAPI int FGAPIENTRY glutStrokeLength(void *font, const unsigned char *string);
/*
* Geometry functions, see freeglut_geometry.c
*/
FGAPI void FGAPIENTRY glutWireCube(GLdouble size);
FGAPI void FGAPIENTRY glutSolidCube(GLdouble size);
FGAPI void FGAPIENTRY glutWireSphere(GLdouble radius, GLint slices, GLint stacks);
FGAPI void FGAPIENTRY glutSolidSphere(GLdouble radius, GLint slices, GLint stacks);
FGAPI void FGAPIENTRY glutWireCone(GLdouble base, GLdouble height, GLint slices, GLint stacks);
FGAPI void FGAPIENTRY glutSolidCone(GLdouble base, GLdouble height, GLint slices, GLint stacks);
FGAPI void FGAPIENTRY glutWireTorus(GLdouble innerRadius, GLdouble outerRadius, GLint sides, GLint rings);
FGAPI void FGAPIENTRY glutSolidTorus(GLdouble innerRadius, GLdouble outerRadius, GLint sides, GLint rings);
FGAPI void FGAPIENTRY glutWireDodecahedron(void);
FGAPI void FGAPIENTRY glutSolidDodecahedron(void);
FGAPI void FGAPIENTRY glutWireOctahedron(void);
FGAPI void FGAPIENTRY glutSolidOctahedron(void);
FGAPI void FGAPIENTRY glutWireTetrahedron(void);
FGAPI void FGAPIENTRY glutSolidTetrahedron(void);
FGAPI void FGAPIENTRY glutWireIcosahedron(void);
FGAPI void FGAPIENTRY glutSolidIcosahedron(void);
/*
* Teapot rendering functions, found in freeglut_teapot.c
*/
FGAPI void FGAPIENTRY glutWireTeapot(GLdouble size);
FGAPI void FGAPIENTRY glutSolidTeapot(GLdouble size);
/*
* Game mode functions, see freeglut_gamemode.c
*/
FGAPI void FGAPIENTRY glutGameModeString(const char *string);
FGAPI int FGAPIENTRY glutEnterGameMode(void);
FGAPI void FGAPIENTRY glutLeaveGameMode(void);
FGAPI int FGAPIENTRY glutGameModeGet(GLenum query);
/*
* Video resize functions, see freeglut_videoresize.c
*/
FGAPI int FGAPIENTRY glutVideoResizeGet(GLenum query);
FGAPI void FGAPIENTRY glutSetupVideoResizing(void);
FGAPI void FGAPIENTRY glutStopVideoResizing(void);
FGAPI void FGAPIENTRY glutVideoResize(int x, int y, int width, int height);
FGAPI void FGAPIENTRY glutVideoPan(int x, int y, int width, int height);
/*
* Colormap functions, see freeglut_misc.c
*/
FGAPI void FGAPIENTRY glutSetColor(int color, GLfloat red, GLfloat green, GLfloat blue);
FGAPI GLfloat FGAPIENTRY glutGetColor(int color, int component);
FGAPI void FGAPIENTRY glutCopyColormap(int window);
/*
* Misc keyboard and joystick functions, see freeglut_misc.c
*/
FGAPI void FGAPIENTRY glutIgnoreKeyRepeat(int ignore);
FGAPI void FGAPIENTRY glutSetKeyRepeat(int repeatMode); /* DEPRECATED 11/4/02 - Do not use */
FGAPI void FGAPIENTRY glutForceJoystickFunc(void);
/*
* Misc functions, see freeglut_misc.c
*/
FGAPI int FGAPIENTRY glutExtensionSupported(const char *extension);
FGAPI void FGAPIENTRY glutReportErrors(void);
#ifdef __cplusplus
}
#endif
/*** END OF FILE ***/
#endif /* __FREEGLUT_STD_H__ */

14457
Common/GL/glew.h Normal file

File diff suppressed because it is too large Load Diff

7125
Common/GL/glext.h Normal file

File diff suppressed because it is too large Load Diff

597
Common/GL/glut.h Normal file
View File

@@ -0,0 +1,597 @@
#ifndef __glut_h__
#define __glut_h__
/* Copyright (c) Mark J. Kilgard, 1994, 1995, 1996, 1998. */
/* This program is freely distributable without licensing fees and is
provided without guarantee or warrantee expressed or implied. This
program is -not- in the public domain. */
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
/* GLUT 3.7 now tries to avoid including <windows.h>
to avoid name space pollution, but Win32's <GL/gl.h>
needs APIENTRY and WINGDIAPI defined properly. */
# if 0
# define WIN32_LEAN_AND_MEAN
# include <windows.h>
# else
/* XXX This is from Win32's <windef.h> */
# ifndef APIENTRY
# define GLUT_APIENTRY_DEFINED
# if (_MSC_VER >= 800) || defined(_STDCALL_SUPPORTED)
# define APIENTRY __stdcall
# else
# define APIENTRY
# endif
# endif
/* XXX This is from Win32's <winnt.h> */
# ifndef CALLBACK
# if (defined(_M_MRX000) || defined(_M_IX86) || defined(_M_ALPHA) || defined(_M_PPC)) && !defined(MIDL_PASS)
# define CALLBACK __stdcall
# else
# define CALLBACK
# endif
# endif
/* XXX This is from Win32's <wingdi.h> and <winnt.h> */
# ifndef WINGDIAPI
# define GLUT_WINGDIAPI_DEFINED
# define WINGDIAPI __declspec(dllimport)
# endif
/* XXX This is from Win32's <ctype.h> */
# ifndef _WCHAR_T_DEFINED
typedef unsigned short wchar_t;
# define _WCHAR_T_DEFINED
# endif
# endif
#pragma comment (lib, "winmm.lib") /* link with Windows MultiMedia lib */
#pragma comment (lib, "opengl32.lib") /* link with Microsoft OpenGL lib */
#pragma comment (lib, "glu32.lib") /* link with OpenGL Utility lib */
#pragma message("Note: including lib: glut32.lib\n")
#pragma comment (lib, "glut32.lib") /* link with Win32 GLUT lib */
#pragma warning (disable:4244) /* Disable bogus conversion warnings. */
#pragma warning (disable:4305) /* VC++ 5.0 version of above warning. */
#endif
#include <GL/gl.h>
#include <GL/glu.h>
/* define APIENTRY and CALLBACK to null string if we aren't on Win32 */
#if !defined(WIN32)
#define APIENTRY
#define GLUT_APIENTRY_DEFINED
#define CALLBACK
#endif
#ifdef __cplusplus
extern "C" {
#endif
/**
GLUT API revision history:
GLUT_API_VERSION is updated to reflect incompatible GLUT
API changes (interface changes, semantic changes, deletions,
or additions).
GLUT_API_VERSION=1 First public release of GLUT. 11/29/94
GLUT_API_VERSION=2 Added support for OpenGL/GLX multisampling,
extension. Supports new input devices like tablet, dial and button
box, and Spaceball. Easy to query OpenGL extensions.
GLUT_API_VERSION=3 glutMenuStatus added.
GLUT_API_VERSION=4 glutInitDisplayString, glutWarpPointer,
glutBitmapLength, glutStrokeLength, glutWindowStatusFunc, dynamic
video resize subAPI, glutPostWindowRedisplay, glutKeyboardUpFunc,
glutSpecialUpFunc, glutIgnoreKeyRepeat, glutSetKeyRepeat,
glutJoystickFunc, glutForceJoystickFunc (NOT FINALIZED!).
**/
#ifndef GLUT_API_VERSION /* allow this to be overriden */
#define GLUT_API_VERSION 3
#endif
/**
GLUT implementation revision history:
GLUT_XLIB_IMPLEMENTATION is updated to reflect both GLUT
API revisions and implementation revisions (ie, bug fixes).
GLUT_XLIB_IMPLEMENTATION=1 mjk's first public release of
GLUT Xlib-based implementation. 11/29/94
GLUT_XLIB_IMPLEMENTATION=2 mjk's second public release of
GLUT Xlib-based implementation providing GLUT version 2
interfaces.
GLUT_XLIB_IMPLEMENTATION=3 mjk's GLUT 2.2 images. 4/17/95
GLUT_XLIB_IMPLEMENTATION=4 mjk's GLUT 2.3 images. 6/?/95
GLUT_XLIB_IMPLEMENTATION=5 mjk's GLUT 3.0 images. 10/?/95
GLUT_XLIB_IMPLEMENTATION=7 mjk's GLUT 3.1+ with glutWarpPoitner. 7/24/96
GLUT_XLIB_IMPLEMENTATION=8 mjk's GLUT 3.1+ with glutWarpPoitner
and video resize. 1/3/97
GLUT_XLIB_IMPLEMENTATION=9 mjk's GLUT 3.4 release with early GLUT 4 routines.
GLUT_XLIB_IMPLEMENTATION=11 Mesa 2.5's GLUT 3.6 release.
GLUT_XLIB_IMPLEMENTATION=12 mjk's GLUT 3.6 release with early GLUT 4 routines + signal handling.
GLUT_XLIB_IMPLEMENTATION=13 mjk's GLUT 3.7 release with GameGLUT support.
**/
#ifndef GLUT_XLIB_IMPLEMENTATION /* Allow this to be overriden. */
#define GLUT_XLIB_IMPLEMENTATION 13
#endif
/* Display mode bit masks. */
#define GLUT_RGB 0
#define GLUT_RGBA GLUT_RGB
#define GLUT_INDEX 1
#define GLUT_SINGLE 0
#define GLUT_DOUBLE 2
#define GLUT_ACCUM 4
#define GLUT_ALPHA 8
#define GLUT_DEPTH 16
#define GLUT_STENCIL 32
#if (GLUT_API_VERSION >= 2)
#define GLUT_MULTISAMPLE 128
#define GLUT_STEREO 256
#endif
#if (GLUT_API_VERSION >= 3)
#define GLUT_LUMINANCE 512
#endif
/* Mouse buttons. */
#define GLUT_LEFT_BUTTON 0
#define GLUT_MIDDLE_BUTTON 1
#define GLUT_RIGHT_BUTTON 2
/* Mouse button state. */
#define GLUT_DOWN 0
#define GLUT_UP 1
#if (GLUT_API_VERSION >= 2)
/* function keys */
#define GLUT_KEY_F1 1
#define GLUT_KEY_F2 2
#define GLUT_KEY_F3 3
#define GLUT_KEY_F4 4
#define GLUT_KEY_F5 5
#define GLUT_KEY_F6 6
#define GLUT_KEY_F7 7
#define GLUT_KEY_F8 8
#define GLUT_KEY_F9 9
#define GLUT_KEY_F10 10
#define GLUT_KEY_F11 11
#define GLUT_KEY_F12 12
/* directional keys */
#define GLUT_KEY_LEFT 100
#define GLUT_KEY_UP 101
#define GLUT_KEY_RIGHT 102
#define GLUT_KEY_DOWN 103
#define GLUT_KEY_PAGE_UP 104
#define GLUT_KEY_PAGE_DOWN 105
#define GLUT_KEY_HOME 106
#define GLUT_KEY_END 107
#define GLUT_KEY_INSERT 108
#endif
/* Entry/exit state. */
#define GLUT_LEFT 0
#define GLUT_ENTERED 1
/* Menu usage state. */
#define GLUT_MENU_NOT_IN_USE 0
#define GLUT_MENU_IN_USE 1
/* Visibility state. */
#define GLUT_NOT_VISIBLE 0
#define GLUT_VISIBLE 1
/* Window status state. */
#define GLUT_HIDDEN 0
#define GLUT_FULLY_RETAINED 1
#define GLUT_PARTIALLY_RETAINED 2
#define GLUT_FULLY_COVERED 3
/* Color index component selection values. */
#define GLUT_RED 0
#define GLUT_GREEN 1
#define GLUT_BLUE 2
/* Layers for use. */
#define GLUT_NORMAL 0
#define GLUT_OVERLAY 1
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
/* Stroke font constants (use these in GLUT program). */
#define GLUT_STROKE_ROMAN ((void*)0)
#define GLUT_STROKE_MONO_ROMAN ((void*)1)
/* Bitmap font constants (use these in GLUT program). */
#define GLUT_BITMAP_9_BY_15 ((void*)2)
#define GLUT_BITMAP_8_BY_13 ((void*)3)
#define GLUT_BITMAP_TIMES_ROMAN_10 ((void*)4)
#define GLUT_BITMAP_TIMES_ROMAN_24 ((void*)5)
#if (GLUT_API_VERSION >= 3)
#define GLUT_BITMAP_HELVETICA_10 ((void*)6)
#define GLUT_BITMAP_HELVETICA_12 ((void*)7)
#define GLUT_BITMAP_HELVETICA_18 ((void*)8)
#endif
#else
/* Stroke font opaque addresses (use constants instead in source code). */
extern void *glutStrokeRoman;
extern void *glutStrokeMonoRoman;
/* Stroke font constants (use these in GLUT program). */
#define GLUT_STROKE_ROMAN (&glutStrokeRoman)
#define GLUT_STROKE_MONO_ROMAN (&glutStrokeMonoRoman)
/* Bitmap font opaque addresses (use constants instead in source code). */
extern void *glutBitmap9By15;
extern void *glutBitmap8By13;
extern void *glutBitmapTimesRoman10;
extern void *glutBitmapTimesRoman24;
extern void *glutBitmapHelvetica10;
extern void *glutBitmapHelvetica12;
extern void *glutBitmapHelvetica18;
/* Bitmap font constants (use these in GLUT program). */
#define GLUT_BITMAP_9_BY_15 (&glutBitmap9By15)
#define GLUT_BITMAP_8_BY_13 (&glutBitmap8By13)
#define GLUT_BITMAP_TIMES_ROMAN_10 (&glutBitmapTimesRoman10)
#define GLUT_BITMAP_TIMES_ROMAN_24 (&glutBitmapTimesRoman24)
#if (GLUT_API_VERSION >= 3)
#define GLUT_BITMAP_HELVETICA_10 (&glutBitmapHelvetica10)
#define GLUT_BITMAP_HELVETICA_12 (&glutBitmapHelvetica12)
#define GLUT_BITMAP_HELVETICA_18 (&glutBitmapHelvetica18)
#endif
#endif
/* glutGet parameters. */
#define GLUT_WINDOW_X 100
#define GLUT_WINDOW_Y 101
#define GLUT_WINDOW_WIDTH 102
#define GLUT_WINDOW_HEIGHT 103
#define GLUT_WINDOW_BUFFER_SIZE 104
#define GLUT_WINDOW_STENCIL_SIZE 105
#define GLUT_WINDOW_DEPTH_SIZE 106
#define GLUT_WINDOW_RED_SIZE 107
#define GLUT_WINDOW_GREEN_SIZE 108
#define GLUT_WINDOW_BLUE_SIZE 109
#define GLUT_WINDOW_ALPHA_SIZE 110
#define GLUT_WINDOW_ACCUM_RED_SIZE 111
#define GLUT_WINDOW_ACCUM_GREEN_SIZE 112
#define GLUT_WINDOW_ACCUM_BLUE_SIZE 113
#define GLUT_WINDOW_ACCUM_ALPHA_SIZE 114
#define GLUT_WINDOW_DOUBLEBUFFER 115
#define GLUT_WINDOW_RGBA 116
#define GLUT_WINDOW_PARENT 117
#define GLUT_WINDOW_NUM_CHILDREN 118
#define GLUT_WINDOW_COLORMAP_SIZE 119
#if (GLUT_API_VERSION >= 2)
#define GLUT_WINDOW_NUM_SAMPLES 120
#define GLUT_WINDOW_STEREO 121
#endif
#if (GLUT_API_VERSION >= 3)
#define GLUT_WINDOW_CURSOR 122
#endif
#define GLUT_SCREEN_WIDTH 200
#define GLUT_SCREEN_HEIGHT 201
#define GLUT_SCREEN_WIDTH_MM 202
#define GLUT_SCREEN_HEIGHT_MM 203
#define GLUT_MENU_NUM_ITEMS 300
#define GLUT_DISPLAY_MODE_POSSIBLE 400
#define GLUT_INIT_WINDOW_X 500
#define GLUT_INIT_WINDOW_Y 501
#define GLUT_INIT_WINDOW_WIDTH 502
#define GLUT_INIT_WINDOW_HEIGHT 503
#define GLUT_INIT_DISPLAY_MODE 504
#if (GLUT_API_VERSION >= 2)
#define GLUT_ELAPSED_TIME 700
#endif
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 13)
#define GLUT_WINDOW_FORMAT_ID 123
#endif
#if (GLUT_API_VERSION >= 2)
/* glutDeviceGet parameters. */
#define GLUT_HAS_KEYBOARD 600
#define GLUT_HAS_MOUSE 601
#define GLUT_HAS_SPACEBALL 602
#define GLUT_HAS_DIAL_AND_BUTTON_BOX 603
#define GLUT_HAS_TABLET 604
#define GLUT_NUM_MOUSE_BUTTONS 605
#define GLUT_NUM_SPACEBALL_BUTTONS 606
#define GLUT_NUM_BUTTON_BOX_BUTTONS 607
#define GLUT_NUM_DIALS 608
#define GLUT_NUM_TABLET_BUTTONS 609
#endif
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 13)
#define GLUT_DEVICE_IGNORE_KEY_REPEAT 610
#define GLUT_DEVICE_KEY_REPEAT 611
#define GLUT_HAS_JOYSTICK 612
#define GLUT_OWNS_JOYSTICK 613
#define GLUT_JOYSTICK_BUTTONS 614
#define GLUT_JOYSTICK_AXES 615
#define GLUT_JOYSTICK_POLL_RATE 616
#endif
#if (GLUT_API_VERSION >= 3)
/* glutLayerGet parameters. */
#define GLUT_OVERLAY_POSSIBLE 800
#define GLUT_LAYER_IN_USE 801
#define GLUT_HAS_OVERLAY 802
#define GLUT_TRANSPARENT_INDEX 803
#define GLUT_NORMAL_DAMAGED 804
#define GLUT_OVERLAY_DAMAGED 805
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
/* glutVideoResizeGet parameters. */
#define GLUT_VIDEO_RESIZE_POSSIBLE 900
#define GLUT_VIDEO_RESIZE_IN_USE 901
#define GLUT_VIDEO_RESIZE_X_DELTA 902
#define GLUT_VIDEO_RESIZE_Y_DELTA 903
#define GLUT_VIDEO_RESIZE_WIDTH_DELTA 904
#define GLUT_VIDEO_RESIZE_HEIGHT_DELTA 905
#define GLUT_VIDEO_RESIZE_X 906
#define GLUT_VIDEO_RESIZE_Y 907
#define GLUT_VIDEO_RESIZE_WIDTH 908
#define GLUT_VIDEO_RESIZE_HEIGHT 909
#endif
/* glutUseLayer parameters. */
#define GLUT_NORMAL 0
#define GLUT_OVERLAY 1
/* glutGetModifiers return mask. */
#define GLUT_ACTIVE_SHIFT 1
#define GLUT_ACTIVE_CTRL 2
#define GLUT_ACTIVE_ALT 4
/* glutSetCursor parameters. */
/* Basic arrows. */
#define GLUT_CURSOR_RIGHT_ARROW 0
#define GLUT_CURSOR_LEFT_ARROW 1
/* Symbolic cursor shapes. */
#define GLUT_CURSOR_INFO 2
#define GLUT_CURSOR_DESTROY 3
#define GLUT_CURSOR_HELP 4
#define GLUT_CURSOR_CYCLE 5
#define GLUT_CURSOR_SPRAY 6
#define GLUT_CURSOR_WAIT 7
#define GLUT_CURSOR_TEXT 8
#define GLUT_CURSOR_CROSSHAIR 9
/* Directional cursors. */
#define GLUT_CURSOR_UP_DOWN 10
#define GLUT_CURSOR_LEFT_RIGHT 11
/* Sizing cursors. */
#define GLUT_CURSOR_TOP_SIDE 12
#define GLUT_CURSOR_BOTTOM_SIDE 13
#define GLUT_CURSOR_LEFT_SIDE 14
#define GLUT_CURSOR_RIGHT_SIDE 15
#define GLUT_CURSOR_TOP_LEFT_CORNER 16
#define GLUT_CURSOR_TOP_RIGHT_CORNER 17
#define GLUT_CURSOR_BOTTOM_RIGHT_CORNER 18
#define GLUT_CURSOR_BOTTOM_LEFT_CORNER 19
/* Inherit from parent window. */
#define GLUT_CURSOR_INHERIT 100
/* Blank cursor. */
#define GLUT_CURSOR_NONE 101
/* Fullscreen crosshair (if available). */
#define GLUT_CURSOR_FULL_CROSSHAIR 102
#endif
/* GLUT initialization sub-API. */
extern void APIENTRY glutInit(int *argcp, char **argv);
extern void APIENTRY glutInitDisplayMode(unsigned int mode);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
extern void APIENTRY glutInitDisplayString(const char *string);
#endif
extern void APIENTRY glutInitWindowPosition(int x, int y);
extern void APIENTRY glutInitWindowSize(int width, int height);
extern void APIENTRY glutMainLoop(void);
/* GLUT window sub-API. */
extern int APIENTRY glutCreateWindow(const char *title);
extern int APIENTRY glutCreateSubWindow(int win, int x, int y, int width, int height);
extern void APIENTRY glutDestroyWindow(int win);
extern void APIENTRY glutPostRedisplay(void);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 11)
extern void APIENTRY glutPostWindowRedisplay(int win);
#endif
extern void APIENTRY glutSwapBuffers(void);
extern int APIENTRY glutGetWindow(void);
extern void APIENTRY glutSetWindow(int win);
extern void APIENTRY glutSetWindowTitle(const char *title);
extern void APIENTRY glutSetIconTitle(const char *title);
extern void APIENTRY glutPositionWindow(int x, int y);
extern void APIENTRY glutReshapeWindow(int width, int height);
extern void APIENTRY glutPopWindow(void);
extern void APIENTRY glutPushWindow(void);
extern void APIENTRY glutIconifyWindow(void);
extern void APIENTRY glutShowWindow(void);
extern void APIENTRY glutHideWindow(void);
#if (GLUT_API_VERSION >= 3)
extern void APIENTRY glutFullScreen(void);
extern void APIENTRY glutSetCursor(int cursor);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
extern void APIENTRY glutWarpPointer(int x, int y);
#endif
/* GLUT overlay sub-API. */
extern void APIENTRY glutEstablishOverlay(void);
extern void APIENTRY glutRemoveOverlay(void);
extern void APIENTRY glutUseLayer(GLenum layer);
extern void APIENTRY glutPostOverlayRedisplay(void);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 11)
extern void APIENTRY glutPostWindowOverlayRedisplay(int win);
#endif
extern void APIENTRY glutShowOverlay(void);
extern void APIENTRY glutHideOverlay(void);
#endif
/* GLUT menu sub-API. */
extern int APIENTRY glutCreateMenu(void ( *)(int));
extern void APIENTRY glutDestroyMenu(int menu);
extern int APIENTRY glutGetMenu(void);
extern void APIENTRY glutSetMenu(int menu);
extern void APIENTRY glutAddMenuEntry(const char *label, int value);
extern void APIENTRY glutAddSubMenu(const char *label, int submenu);
extern void APIENTRY glutChangeToMenuEntry(int item, const char *label, int value);
extern void APIENTRY glutChangeToSubMenu(int item, const char *label, int submenu);
extern void APIENTRY glutRemoveMenuItem(int item);
extern void APIENTRY glutAttachMenu(int button);
extern void APIENTRY glutDetachMenu(int button);
/* GLUT window callback sub-API. */
extern void APIENTRY glutDisplayFunc(void (*func)(void));
extern void APIENTRY glutReshapeFunc(void (*func)(int width, int height));
extern void APIENTRY glutKeyboardFunc(void (*func)(unsigned char key, int x, int y));
extern void APIENTRY glutMouseFunc(void (*func)(int button, int state, int x, int y));
extern void APIENTRY glutMotionFunc(void (*func)(int x, int y));
extern void APIENTRY glutPassiveMotionFunc(void (*func)(int x, int y));
extern void APIENTRY glutEntryFunc(void (*func)(int state));
extern void APIENTRY glutVisibilityFunc(void (*func)(int state));
extern void APIENTRY glutIdleFunc(void (*func)(void));
extern void APIENTRY glutTimerFunc(unsigned int millis, void (*func)(int value), int value);
extern void APIENTRY glutMenuStateFunc(void (*func)(int state));
#if (GLUT_API_VERSION >= 2)
extern void APIENTRY glutSpecialFunc(void (*func)(int key, int x, int y));
extern void APIENTRY glutSpaceballMotionFunc(void (*func)(int x, int y, int z));
extern void APIENTRY glutSpaceballRotateFunc(void (*func)(int x, int y, int z));
extern void APIENTRY glutSpaceballButtonFunc(void (*func)(int button, int state));
extern void APIENTRY glutButtonBoxFunc(void (*func)(int button, int state));
extern void APIENTRY glutDialsFunc(void (*func)(int dial, int value));
extern void APIENTRY glutTabletMotionFunc(void (*func)(int x, int y));
extern void APIENTRY glutTabletButtonFunc(void (*func)(int button, int state, int x, int y));
#if (GLUT_API_VERSION >= 3)
extern void APIENTRY glutMenuStatusFunc(void (*func)(int status, int x, int y));
extern void APIENTRY glutOverlayDisplayFunc(void (*func)(void));
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
extern void APIENTRY glutWindowStatusFunc(void (*func)(int state));
#endif
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 13)
extern void APIENTRY glutKeyboardUpFunc(void (*func)(unsigned char key, int x, int y));
extern void APIENTRY glutSpecialUpFunc(void (*func)(int key, int x, int y));
extern void APIENTRY glutJoystickFunc(void (*func)(unsigned int buttonMask, int x, int y, int z), int pollInterval);
#endif
#endif
#endif
/* GLUT color index sub-API. */
extern void APIENTRY glutSetColor(int, GLfloat red, GLfloat green, GLfloat blue);
extern GLfloat APIENTRY glutGetColor(int ndx, int component);
extern void APIENTRY glutCopyColormap(int win);
/* GLUT state retrieval sub-API. */
extern int APIENTRY glutGet(GLenum type);
extern int APIENTRY glutDeviceGet(GLenum type);
#if (GLUT_API_VERSION >= 2)
/* GLUT extension support sub-API */
extern int APIENTRY glutExtensionSupported(const char *name);
#endif
#if (GLUT_API_VERSION >= 3)
extern int APIENTRY glutGetModifiers(void);
extern int APIENTRY glutLayerGet(GLenum type);
#endif
/* GLUT font sub-API */
extern void APIENTRY glutBitmapCharacter(void *font, int character);
extern int APIENTRY glutBitmapWidth(void *font, int character);
extern void APIENTRY glutStrokeCharacter(void *font, int character);
extern int APIENTRY glutStrokeWidth(void *font, int character);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
extern int APIENTRY glutBitmapLength(void *font, const unsigned char *string);
extern int APIENTRY glutStrokeLength(void *font, const unsigned char *string);
#endif
/* GLUT pre-built models sub-API */
extern void APIENTRY glutWireSphere(GLdouble radius, GLint slices, GLint stacks);
extern void APIENTRY glutSolidSphere(GLdouble radius, GLint slices, GLint stacks);
extern void APIENTRY glutWireCone(GLdouble base, GLdouble height, GLint slices, GLint stacks);
extern void APIENTRY glutSolidCone(GLdouble base, GLdouble height, GLint slices, GLint stacks);
extern void APIENTRY glutWireCube(GLdouble size);
extern void APIENTRY glutSolidCube(GLdouble size);
extern void APIENTRY glutWireTorus(GLdouble innerRadius, GLdouble outerRadius, GLint sides, GLint rings);
extern void APIENTRY glutSolidTorus(GLdouble innerRadius, GLdouble outerRadius, GLint sides, GLint rings);
extern void APIENTRY glutWireDodecahedron(void);
extern void APIENTRY glutSolidDodecahedron(void);
extern void APIENTRY glutWireTeapot(GLdouble size);
extern void APIENTRY glutSolidTeapot(GLdouble size);
extern void APIENTRY glutWireOctahedron(void);
extern void APIENTRY glutSolidOctahedron(void);
extern void APIENTRY glutWireTetrahedron(void);
extern void APIENTRY glutSolidTetrahedron(void);
extern void APIENTRY glutWireIcosahedron(void);
extern void APIENTRY glutSolidIcosahedron(void);
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 9)
/* GLUT video resize sub-API. */
extern int APIENTRY glutVideoResizeGet(GLenum param);
extern void APIENTRY glutSetupVideoResizing(void);
extern void APIENTRY glutStopVideoResizing(void);
extern void APIENTRY glutVideoResize(int x, int y, int width, int height);
extern void APIENTRY glutVideoPan(int x, int y, int width, int height);
/* GLUT debugging sub-API. */
extern void APIENTRY glutReportErrors(void);
#endif
#if (GLUT_API_VERSION >= 4 || GLUT_XLIB_IMPLEMENTATION >= 13)
/* GLUT device control sub-API. */
/* glutSetKeyRepeat modes. */
#define GLUT_KEY_REPEAT_OFF 0
#define GLUT_KEY_REPEAT_ON 1
#define GLUT_KEY_REPEAT_DEFAULT 2
/* Joystick button masks. */
#define GLUT_JOYSTICK_BUTTON_A 1
#define GLUT_JOYSTICK_BUTTON_B 2
#define GLUT_JOYSTICK_BUTTON_C 4
#define GLUT_JOYSTICK_BUTTON_D 8
extern void APIENTRY glutIgnoreKeyRepeat(int ignore);
extern void APIENTRY glutSetKeyRepeat(int repeatMode);
extern void APIENTRY glutForceJoystickFunc(void);
/* GLUT game mode sub-API. */
/* glutGameModeGet. */
#define GLUT_GAME_MODE_ACTIVE 0
#define GLUT_GAME_MODE_POSSIBLE 1
#define GLUT_GAME_MODE_WIDTH 2
#define GLUT_GAME_MODE_HEIGHT 3
#define GLUT_GAME_MODE_PIXEL_DEPTH 4
#define GLUT_GAME_MODE_REFRESH_RATE 5
#define GLUT_GAME_MODE_DISPLAY_CHANGED 6
extern void APIENTRY glutGameModeString(const char *string);
extern int APIENTRY glutEnterGameMode(void);
extern void APIENTRY glutLeaveGameMode(void);
extern int APIENTRY glutGameModeGet(GLenum mode);
#endif
#ifdef __cplusplus
}
#endif
#ifdef GLUT_APIENTRY_DEFINED
# undef GLUT_APIENTRY_DEFINED
# undef APIENTRY
#endif
#ifdef GLUT_WINGDIAPI_DEFINED
# undef GLUT_WINGDIAPI_DEFINED
# undef WINGDIAPI
#endif
#endif /* __glut_h__ */

1121
Common/GL/glxew.h Normal file

File diff suppressed because it is too large Load Diff

805
Common/GL/glxext.h Normal file
View File

@@ -0,0 +1,805 @@
#ifndef __glxext_h_
#define __glxext_h_
#ifdef __cplusplus
extern "C" {
#endif
/*
** License Applicability. Except to the extent portions of this file are
** made subject to an alternative license as permitted in the SGI Free
** Software License B, Version 1.1 (the "License"), the contents of this
** file are subject only to the provisions of the License. You may not use
** this file except in compliance with the License. You may obtain a copy
** of the License at Silicon Graphics, Inc., attn: Legal Services, 1600
** Amphitheatre Parkway, Mountain View, CA 94043-1351, or at:
**
** http://oss.sgi.com/projects/FreeB
**
** Note that, as provided in the License, the Software is distributed on an
** "AS IS" basis, with ALL EXPRESS AND IMPLIED WARRANTIES AND CONDITIONS
** DISCLAIMED, INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTIES AND
** CONDITIONS OF MERCHANTABILITY, SATISFACTORY QUALITY, FITNESS FOR A
** PARTICULAR PURPOSE, AND NON-INFRINGEMENT.
**
** Original Code. The Original Code is: OpenGL Sample Implementation,
** Version 1.2.1, released January 26, 2000, developed by Silicon Graphics,
** Inc. The Original Code is Copyright (c) 1991-2004 Silicon Graphics, Inc.
** Copyright in any portions created by third parties is as indicated
** elsewhere herein. All Rights Reserved.
**
** Additional Notice Provisions: This software was created using the
** OpenGL(R) version 1.2.1 Sample Implementation published by SGI, but has
** not been independently verified as being compliant with the OpenGL(R)
** version 1.2.1 Specification.
*/
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64) && !defined(APIENTRY) && !defined(__CYGWIN__) && !defined(__SCITECH_SNAP__)
#define WIN32_LEAN_AND_MEAN 1
#include <windows.h>
#endif
#ifndef APIENTRY
#define APIENTRY
#endif
#ifndef APIENTRYP
#define APIENTRYP APIENTRY *
#endif
#ifndef GLAPI
#define GLAPI extern
#endif
/*************************************************************/
/* Header file version number, required by OpenGL ABI for Linux */
/* glxext.h last updated 2005/01/20 */
/* Current version at http://oss.sgi.com/projects/ogl-sample/registry/ */
#define GLX_GLXEXT_VERSION 10
#ifndef GLX_ARB_get_proc_address
#endif
#ifndef GLX_ARB_multisample
#define GLX_SAMPLE_BUFFERS_ARB 100000
#define GLX_SAMPLES_ARB 100001
#endif
#ifndef GLX_ARB_fbconfig_float
#define GLX_RGBA_FLOAT_TYPE_ARB 0x20B9
#define GLX_RGBA_FLOAT_BIT_ARB 0x00000004
#endif
#ifndef GLX_SGIS_multisample
#define GLX_SAMPLE_BUFFERS_SGIS 100000
#define GLX_SAMPLES_SGIS 100001
#endif
#ifndef GLX_EXT_visual_info
#define GLX_X_VISUAL_TYPE_EXT 0x22
#define GLX_TRANSPARENT_TYPE_EXT 0x23
#define GLX_TRANSPARENT_INDEX_VALUE_EXT 0x24
#define GLX_TRANSPARENT_RED_VALUE_EXT 0x25
#define GLX_TRANSPARENT_GREEN_VALUE_EXT 0x26
#define GLX_TRANSPARENT_BLUE_VALUE_EXT 0x27
#define GLX_TRANSPARENT_ALPHA_VALUE_EXT 0x28
#define GLX_NONE_EXT 0x8000
#define GLX_TRUE_COLOR_EXT 0x8002
#define GLX_DIRECT_COLOR_EXT 0x8003
#define GLX_PSEUDO_COLOR_EXT 0x8004
#define GLX_STATIC_COLOR_EXT 0x8005
#define GLX_GRAY_SCALE_EXT 0x8006
#define GLX_STATIC_GRAY_EXT 0x8007
#define GLX_TRANSPARENT_RGB_EXT 0x8008
#define GLX_TRANSPARENT_INDEX_EXT 0x8009
#endif
#ifndef GLX_SGI_swap_control
#endif
#ifndef GLX_SGI_video_sync
#endif
#ifndef GLX_SGI_make_current_read
#endif
#ifndef GLX_SGIX_video_source
#endif
#ifndef GLX_EXT_visual_rating
#define GLX_VISUAL_CAVEAT_EXT 0x20
#define GLX_SLOW_VISUAL_EXT 0x8001
#define GLX_NON_CONFORMANT_VISUAL_EXT 0x800D
/* reuse GLX_NONE_EXT */
#endif
#ifndef GLX_EXT_import_context
#define GLX_SHARE_CONTEXT_EXT 0x800A
#define GLX_VISUAL_ID_EXT 0x800B
#define GLX_SCREEN_EXT 0x800C
#endif
#ifndef GLX_SGIX_fbconfig
#define GLX_WINDOW_BIT_SGIX 0x00000001
#define GLX_PIXMAP_BIT_SGIX 0x00000002
#define GLX_RGBA_BIT_SGIX 0x00000001
#define GLX_COLOR_INDEX_BIT_SGIX 0x00000002
#define GLX_DRAWABLE_TYPE_SGIX 0x8010
#define GLX_RENDER_TYPE_SGIX 0x8011
#define GLX_X_RENDERABLE_SGIX 0x8012
#define GLX_FBCONFIG_ID_SGIX 0x8013
#define GLX_RGBA_TYPE_SGIX 0x8014
#define GLX_COLOR_INDEX_TYPE_SGIX 0x8015
/* reuse GLX_SCREEN_EXT */
#endif
#ifndef GLX_SGIX_pbuffer
#define GLX_PBUFFER_BIT_SGIX 0x00000004
#define GLX_BUFFER_CLOBBER_MASK_SGIX 0x08000000
#define GLX_FRONT_LEFT_BUFFER_BIT_SGIX 0x00000001
#define GLX_FRONT_RIGHT_BUFFER_BIT_SGIX 0x00000002
#define GLX_BACK_LEFT_BUFFER_BIT_SGIX 0x00000004
#define GLX_BACK_RIGHT_BUFFER_BIT_SGIX 0x00000008
#define GLX_AUX_BUFFERS_BIT_SGIX 0x00000010
#define GLX_DEPTH_BUFFER_BIT_SGIX 0x00000020
#define GLX_STENCIL_BUFFER_BIT_SGIX 0x00000040
#define GLX_ACCUM_BUFFER_BIT_SGIX 0x00000080
#define GLX_SAMPLE_BUFFERS_BIT_SGIX 0x00000100
#define GLX_MAX_PBUFFER_WIDTH_SGIX 0x8016
#define GLX_MAX_PBUFFER_HEIGHT_SGIX 0x8017
#define GLX_MAX_PBUFFER_PIXELS_SGIX 0x8018
#define GLX_OPTIMAL_PBUFFER_WIDTH_SGIX 0x8019
#define GLX_OPTIMAL_PBUFFER_HEIGHT_SGIX 0x801A
#define GLX_PRESERVED_CONTENTS_SGIX 0x801B
#define GLX_LARGEST_PBUFFER_SGIX 0x801C
#define GLX_WIDTH_SGIX 0x801D
#define GLX_HEIGHT_SGIX 0x801E
#define GLX_EVENT_MASK_SGIX 0x801F
#define GLX_DAMAGED_SGIX 0x8020
#define GLX_SAVED_SGIX 0x8021
#define GLX_WINDOW_SGIX 0x8022
#define GLX_PBUFFER_SGIX 0x8023
#endif
#ifndef GLX_SGI_cushion
#endif
#ifndef GLX_SGIX_video_resize
#define GLX_SYNC_FRAME_SGIX 0x00000000
#define GLX_SYNC_SWAP_SGIX 0x00000001
#endif
#ifndef GLX_SGIX_dmbuffer
#define GLX_DIGITAL_MEDIA_PBUFFER_SGIX 0x8024
#endif
#ifndef GLX_SGIX_swap_group
#endif
#ifndef GLX_SGIX_swap_barrier
#endif
#ifndef GLX_SGIS_blended_overlay
#define GLX_BLENDED_RGBA_SGIS 0x8025
#endif
#ifndef GLX_SGIS_shared_multisample
#define GLX_MULTISAMPLE_SUB_RECT_WIDTH_SGIS 0x8026
#define GLX_MULTISAMPLE_SUB_RECT_HEIGHT_SGIS 0x8027
#endif
#ifndef GLX_SUN_get_transparent_index
#endif
#ifndef GLX_3DFX_multisample
#define GLX_SAMPLE_BUFFERS_3DFX 0x8050
#define GLX_SAMPLES_3DFX 0x8051
#endif
#ifndef GLX_MESA_copy_sub_buffer
#endif
#ifndef GLX_MESA_pixmap_colormap
#endif
#ifndef GLX_MESA_release_buffers
#endif
#ifndef GLX_MESA_set_3dfx_mode
#define GLX_3DFX_WINDOW_MODE_MESA 0x1
#define GLX_3DFX_FULLSCREEN_MODE_MESA 0x2
#endif
#ifndef GLX_SGIX_visual_select_group
#define GLX_VISUAL_SELECT_GROUP_SGIX 0x8028
#endif
#ifndef GLX_OML_swap_method
#define GLX_SWAP_METHOD_OML 0x8060
#define GLX_SWAP_EXCHANGE_OML 0x8061
#define GLX_SWAP_COPY_OML 0x8062
#define GLX_SWAP_UNDEFINED_OML 0x8063
#endif
#ifndef GLX_OML_sync_control
#endif
#ifndef GLX_NV_float_buffer
#define GLX_FLOAT_COMPONENTS_NV 0x20B0
#endif
#ifndef GLX_SGIX_hyperpipe
#define GLX_HYPERPIPE_PIPE_NAME_LENGTH_SGIX 80
#define GLX_BAD_HYPERPIPE_CONFIG_SGIX 91
#define GLX_BAD_HYPERPIPE_SGIX 92
#define GLX_HYPERPIPE_DISPLAY_PIPE_SGIX 0x00000001
#define GLX_HYPERPIPE_RENDER_PIPE_SGIX 0x00000002
#define GLX_PIPE_RECT_SGIX 0x00000001
#define GLX_PIPE_RECT_LIMITS_SGIX 0x00000002
#define GLX_HYPERPIPE_STEREO_SGIX 0x00000003
#define GLX_HYPERPIPE_PIXEL_AVERAGE_SGIX 0x00000004
#define GLX_HYPERPIPE_ID_SGIX 0x8030
#endif
#ifndef GLX_MESA_agp_offset
#endif
/*************************************************************/
#ifndef GLX_ARB_get_proc_address
/*
* Linux OpenGL ABI specifies glXGetProcAddressARB should be
* in glx.h moving related defines there as well.
*/
#endif
#ifndef GLX_SGIX_video_source
typedef XID GLXVideoSourceSGIX;
#endif
#ifndef GLX_SGIX_fbconfig
typedef XID GLXFBConfigIDSGIX;
typedef struct __GLXFBConfigRec *GLXFBConfigSGIX;
#endif
#ifndef GLX_SGIX_pbuffer
typedef XID GLXPbufferSGIX;
typedef struct
{
int type;
unsigned long serial; /* # of last request processed by server */
Bool send_event; /* true if this came for SendEvent request */
Display *display; /* display the event was read from */
GLXDrawable drawable; /* i.d. of Drawable */
int event_type; /* GLX_DAMAGED_SGIX or GLX_SAVED_SGIX */
int draw_type; /* GLX_WINDOW_SGIX or GLX_PBUFFER_SGIX */
unsigned int mask; /* mask indicating which buffers are affected*/
int x, y;
int width, height;
int count; /* if nonzero, at least this many more */
} GLXBufferClobberEventSGIX;
#endif
#ifndef GLX_NV_swap_group
#endif
#ifndef GLX_NV_video_out
/*
* GLXVideoDeviceNV is an opaque handle to a video device (part of the
* GLX_NV_video_out extension).
*/
typedef unsigned int GLXVideoDeviceNV;
/* glXBindVideoImageNV iVideoBuffer values (NV_video_out) */
#define GLX_VIDEO_OUT_COLOR_NV 0x20C3
#define GLX_VIDEO_OUT_ALPHA_NV 0x20C4
#define GLX_VIDEO_OUT_DEPTH_NV 0x20C5
#define GLX_VIDEO_OUT_COLOR_AND_ALPHA_NV 0x20C6
#define GLX_VIDEO_OUT_COLOR_AND_DEPTH_NV 0x20C7
/* glXSendPbufferToVideoNV iBufferType values (NV_video_out) */
#define GLX_VIDEO_OUT_FRAME_NV 0x20C8
#define GLX_VIDEO_OUT_FIELD_1_NV 0x20C9
#define GLX_VIDEO_OUT_FIELD_2_NV 0x20CA
#endif
#ifndef GLX_EXT_texture_from_pixmap
/* New glXGetFBConfigAttrib <attrib_list> tokens */
#define GLX_BIND_TO_TEXTURE_RGB_EXT 0x20D0
#define GLX_BIND_TO_TEXTURE_RGBA_EXT 0x20D1
#define GLX_BIND_TO_MIPMAP_TEXTURE_EXT 0x20D2
#define GLX_BIND_TO_TEXTURE_TARGETS_EXT 0x20D3
#define GLX_Y_INVERTED_EXT 0x20D4
/* New glXCreatePixmap attributes and glXQueryDrawable attributes */
#define GLX_TEXTURE_FORMAT_EXT 0x20D5
#define GLX_TEXTURE_TARGET_EXT 0x20D6
#define GLX_MIPMAP_TEXTURE_EXT 0x20D7
/* Values for GLX_TEXTURE_FORMAT_EXT */
#define GLX_TEXTURE_FORMAT_NONE_EXT 0x20D8
#define GLX_TEXTURE_FORMAT_RGB_EXT 0x20D9
#define GLX_TEXTURE_FORMAT_RGBA_EXT 0x20DA
/* Bits for GLX_BIND_TO_TEXTURE_TARGETS_EXT mask */
#define GLX_TEXTURE_1D_BIT_EXT 0x00000001
#define GLX_TEXTURE_2D_BIT_EXT 0x00000002
#define GLX_TEXTURE_RECTANGLE_BIT_EXT 0x00000004
/* Values for GLX_TEXTURE_TARGET_EXT */
#define GLX_TEXTURE_1D_EXT 0x20DB
#define GLX_TEXTURE_2D_EXT 0x20DC
#define GLX_TEXTURE_RECTANGLE_EXT 0x20DD
/*
* Values for the buffer parameter of glXBindTexImageEXT and
* glXReleaseTexImageEXT
*/
#define GLX_FRONT_LEFT_EXT 0x20DE
#define GLX_FRONT_RIGHT_EXT 0x20DF
#define GLX_BACK_LEFT_EXT 0x20E0
#define GLX_BACK_RIGHT_EXT 0x20E1
#define GLX_FRONT_EXT GLX_FRONT_LEFT_EXT
#define GLX_BACK_EXT GLX_BACK_LEFT_EXT
#define GLX_AUX0_EXT 0x20E2
#define GLX_AUX1_EXT 0x20E3
#define GLX_AUX2_EXT 0x20E4
#define GLX_AUX3_EXT 0x20E5
#define GLX_AUX4_EXT 0x20E6
#define GLX_AUX5_EXT 0x20E7
#define GLX_AUX6_EXT 0x20E8
#define GLX_AUX7_EXT 0x20E9
#define GLX_AUX8_EXT 0x20EA
#define GLX_AUX9_EXT 0x20EB
#endif
/* Define int32_t and int64_t types for UST/MSC */
/* (as used in the GLX_OML_sync_control extension). */
#if defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L
#include <inttypes.h>
#elif defined( __VMS ) || defined(__FreeBSD__)
#include <inttypes.h>
#elif (defined(__sun__) && defined(__svr4__)) || (defined(__sun) && defined(__SVR4))
#include <inttypes.h>
#elif defined(__SCO__) || defined(__USLC__) || defined(__linux__)
#include <stdint.h>
#elif defined(__UNIXOS2__) || defined(__SOL64__)
typedef long int int32_t;
typedef long long int int64_t;
#else
#error "int32_t and int64_t are undefined!"
#endif
#ifndef GLX_ARB_get_proc_address
/* Moved to glx.h */
#endif
#ifndef GLX_ARB_multisample
#define GLX_ARB_multisample 1
#endif
#ifndef GLX_ARB_fbconfig_float
#define GLX_ARB_fbconfig_float 1
#endif
#ifndef GLX_SGIS_multisample
#define GLX_SGIS_multisample 1
#endif
#ifndef GLX_EXT_visual_info
#define GLX_EXT_visual_info 1
#endif
#ifndef GLX_SGI_swap_control
#define GLX_SGI_swap_control 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern int glXSwapIntervalSGI(int);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef int (* PFNGLXSWAPINTERVALSGIPROC)(int interval);
#endif
#ifndef GLX_SGI_video_sync
#define GLX_SGI_video_sync 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern int glXGetVideoSyncSGI(unsigned int *);
extern int glXWaitVideoSyncSGI(int, int, unsigned int *);
extern int glXGetRefreshRateSGI(unsigned int *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef int (* PFNGLXGETVIDEOSYNCSGIPROC)(unsigned int *count);
typedef int (* PFNGLXWAITVIDEOSYNCSGIPROC)(int divisor, int remainder, unsigned int *count);
typedef int (* PFNGLXGETREFRESHRATESGIPROC)(unsigned int *);
#endif
#ifndef GLX_SGI_make_current_read
#define GLX_SGI_make_current_read 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXMakeCurrentReadSGI(Display *, GLXDrawable, GLXDrawable, GLXContext);
extern GLXDrawable glXGetCurrentReadDrawableSGI(void);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Bool(* PFNGLXMAKECURRENTREADSGIPROC)(Display *dpy, GLXDrawable draw, GLXDrawable read, GLXContext ctx);
typedef GLXDrawable(* PFNGLXGETCURRENTREADDRAWABLESGIPROC)(void);
#endif
#ifndef GLX_SGIX_video_source
#define GLX_SGIX_video_source 1
#ifdef _VL_H
#ifdef GLX_GLXEXT_PROTOTYPES
extern GLXVideoSourceSGIX glXCreateGLXVideoSourceSGIX(Display *, int, VLServer, VLPath, int, VLNode);
extern void glXDestroyGLXVideoSourceSGIX(Display *, GLXVideoSourceSGIX);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef GLXVideoSourceSGIX(* PFNGLXCREATEGLXVIDEOSOURCESGIXPROC)(Display *display, int screen, VLServer server, VLPath path, int nodeClass, VLNode drainNode);
typedef void (* PFNGLXDESTROYGLXVIDEOSOURCESGIXPROC)(Display *dpy, GLXVideoSourceSGIX glxvideosource);
#endif /* _VL_H */
#endif
#ifndef GLX_EXT_visual_rating
#define GLX_EXT_visual_rating 1
#endif
#ifndef GLX_EXT_import_context
#define GLX_EXT_import_context 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Display *glXGetCurrentDisplayEXT(void);
extern int glXQueryContextInfoEXT(Display *, GLXContext, int, int *);
extern GLXContextID glXGetContextIDEXT(const GLXContext);
extern GLXContext glXImportContextEXT(Display *, GLXContextID);
extern void glXFreeContextEXT(Display *, GLXContext);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Display *(* PFNGLXGETCURRENTDISPLAYEXTPROC)(void);
typedef int (* PFNGLXQUERYCONTEXTINFOEXTPROC)(Display *dpy, GLXContext context, int attribute, int *value);
typedef GLXContextID(* PFNGLXGETCONTEXTIDEXTPROC)(const GLXContext context);
typedef GLXContext(* PFNGLXIMPORTCONTEXTEXTPROC)(Display *dpy, GLXContextID contextID);
typedef void (* PFNGLXFREECONTEXTEXTPROC)(Display *dpy, GLXContext context);
#endif
#ifndef GLX_SGIX_fbconfig
#define GLX_SGIX_fbconfig 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern int glXGetFBConfigAttribSGIX(Display *, GLXFBConfigSGIX, int, int *);
extern GLXFBConfigSGIX *glXChooseFBConfigSGIX(Display *, int, int *, int *);
extern GLXPixmap glXCreateGLXPixmapWithConfigSGIX(Display *, GLXFBConfigSGIX, Pixmap);
extern GLXContext glXCreateContextWithConfigSGIX(Display *, GLXFBConfigSGIX, int, GLXContext, Bool);
extern XVisualInfo *glXGetVisualFromFBConfigSGIX(Display *, GLXFBConfigSGIX);
extern GLXFBConfigSGIX glXGetFBConfigFromVisualSGIX(Display *, XVisualInfo *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef int (* PFNGLXGETFBCONFIGATTRIBSGIXPROC)(Display *dpy, GLXFBConfigSGIX config, int attribute, int *value);
typedef GLXFBConfigSGIX *(* PFNGLXCHOOSEFBCONFIGSGIXPROC)(Display *dpy, int screen, int *attrib_list, int *nelements);
typedef GLXPixmap(* PFNGLXCREATEGLXPIXMAPWITHCONFIGSGIXPROC)(Display *dpy, GLXFBConfigSGIX config, Pixmap pixmap);
typedef GLXContext(* PFNGLXCREATECONTEXTWITHCONFIGSGIXPROC)(Display *dpy, GLXFBConfigSGIX config, int render_type, GLXContext share_list, Bool direct);
typedef XVisualInfo *(* PFNGLXGETVISUALFROMFBCONFIGSGIXPROC)(Display *dpy, GLXFBConfigSGIX config);
typedef GLXFBConfigSGIX(* PFNGLXGETFBCONFIGFROMVISUALSGIXPROC)(Display *dpy, XVisualInfo *vis);
#endif
#ifndef GLX_SGIX_pbuffer
#define GLX_SGIX_pbuffer 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern GLXPbufferSGIX glXCreateGLXPbufferSGIX(Display *, GLXFBConfigSGIX, unsigned int, unsigned int, int *);
extern void glXDestroyGLXPbufferSGIX(Display *, GLXPbufferSGIX);
extern int glXQueryGLXPbufferSGIX(Display *, GLXPbufferSGIX, int, unsigned int *);
extern void glXSelectEventSGIX(Display *, GLXDrawable, unsigned long);
extern void glXGetSelectedEventSGIX(Display *, GLXDrawable, unsigned long *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef GLXPbufferSGIX(* PFNGLXCREATEGLXPBUFFERSGIXPROC)(Display *dpy, GLXFBConfigSGIX config, unsigned int width, unsigned int height, int *attrib_list);
typedef void (* PFNGLXDESTROYGLXPBUFFERSGIXPROC)(Display *dpy, GLXPbufferSGIX pbuf);
typedef int (* PFNGLXQUERYGLXPBUFFERSGIXPROC)(Display *dpy, GLXPbufferSGIX pbuf, int attribute, unsigned int *value);
typedef void (* PFNGLXSELECTEVENTSGIXPROC)(Display *dpy, GLXDrawable drawable, unsigned long mask);
typedef void (* PFNGLXGETSELECTEDEVENTSGIXPROC)(Display *dpy, GLXDrawable drawable, unsigned long *mask);
#endif
#ifndef GLX_SGI_cushion
#define GLX_SGI_cushion 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern void glXCushionSGI(Display *, Window, float);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef void (* PFNGLXCUSHIONSGIPROC)(Display *dpy, Window window, float cushion);
#endif
#ifndef GLX_SGIX_video_resize
#define GLX_SGIX_video_resize 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern int glXBindChannelToWindowSGIX(Display *, int, int, Window);
extern int glXChannelRectSGIX(Display *, int, int, int, int, int, int);
extern int glXQueryChannelRectSGIX(Display *, int, int, int *, int *, int *, int *);
extern int glXQueryChannelDeltasSGIX(Display *, int, int, int *, int *, int *, int *);
extern int glXChannelRectSyncSGIX(Display *, int, int, GLenum);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef int (* PFNGLXBINDCHANNELTOWINDOWSGIXPROC)(Display *display, int screen, int channel, Window window);
typedef int (* PFNGLXCHANNELRECTSGIXPROC)(Display *display, int screen, int channel, int x, int y, int w, int h);
typedef int (* PFNGLXQUERYCHANNELRECTSGIXPROC)(Display *display, int screen, int channel, int *dx, int *dy, int *dw, int *dh);
typedef int (* PFNGLXQUERYCHANNELDELTASSGIXPROC)(Display *display, int screen, int channel, int *x, int *y, int *w, int *h);
typedef int (* PFNGLXCHANNELRECTSYNCSGIXPROC)(Display *display, int screen, int channel, GLenum synctype);
#endif
#ifndef GLX_SGIX_dmbuffer
#define GLX_SGIX_dmbuffer 1
#ifdef _DM_BUFFER_H_
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXAssociateDMPbufferSGIX(Display *, GLXPbufferSGIX, DMparams *, DMbuffer);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Bool(* PFNGLXASSOCIATEDMPBUFFERSGIXPROC)(Display *dpy, GLXPbufferSGIX pbuffer, DMparams *params, DMbuffer dmbuffer);
#endif /* _DM_BUFFER_H_ */
#endif
#ifndef GLX_SGIX_swap_group
#define GLX_SGIX_swap_group 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern void glXJoinSwapGroupSGIX(Display *, GLXDrawable, GLXDrawable);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef void (* PFNGLXJOINSWAPGROUPSGIXPROC)(Display *dpy, GLXDrawable drawable, GLXDrawable member);
#endif
#ifndef GLX_SGIX_swap_barrier
#define GLX_SGIX_swap_barrier 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern void glXBindSwapBarrierSGIX(Display *, GLXDrawable, int);
extern Bool glXQueryMaxSwapBarriersSGIX(Display *, int, int *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef void (* PFNGLXBINDSWAPBARRIERSGIXPROC)(Display *dpy, GLXDrawable drawable, int barrier);
typedef Bool(* PFNGLXQUERYMAXSWAPBARRIERSSGIXPROC)(Display *dpy, int screen, int *max);
#endif
#ifndef GLX_SUN_get_transparent_index
#define GLX_SUN_get_transparent_index 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Status glXGetTransparentIndexSUN(Display *, Window, Window, long *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Status(* PFNGLXGETTRANSPARENTINDEXSUNPROC)(Display *dpy, Window overlay, Window underlay, long *pTransparentIndex);
#endif
#ifndef GLX_MESA_copy_sub_buffer
#define GLX_MESA_copy_sub_buffer 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern void glXCopySubBufferMESA(Display *, GLXDrawable, int, int, int, int);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef void (* PFNGLXCOPYSUBBUFFERMESAPROC)(Display *dpy, GLXDrawable drawable, int x, int y, int width, int height);
#endif
#ifndef GLX_MESA_pixmap_colormap
#define GLX_MESA_pixmap_colormap 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern GLXPixmap glXCreateGLXPixmapMESA(Display *, XVisualInfo *, Pixmap, Colormap);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef GLXPixmap(* PFNGLXCREATEGLXPIXMAPMESAPROC)(Display *dpy, XVisualInfo *visual, Pixmap pixmap, Colormap cmap);
#endif
#ifndef GLX_MESA_release_buffers
#define GLX_MESA_release_buffers 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXReleaseBuffersMESA(Display *, GLXDrawable);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Bool(* PFNGLXRELEASEBUFFERSMESAPROC)(Display *dpy, GLXDrawable drawable);
#endif
#ifndef GLX_MESA_set_3dfx_mode
#define GLX_MESA_set_3dfx_mode 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXSet3DfxModeMESA(int);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Bool(* PFNGLXSET3DFXMODEMESAPROC)(int mode);
#endif
#ifndef GLX_SGIX_visual_select_group
#define GLX_SGIX_visual_select_group 1
#endif
#ifndef GLX_OML_swap_method
#define GLX_OML_swap_method 1
#endif
#ifndef GLX_OML_sync_control
#define GLX_OML_sync_control 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXGetSyncValuesOML(Display *, GLXDrawable, int64_t *, int64_t *, int64_t *);
extern Bool glXGetMscRateOML(Display *, GLXDrawable, int32_t *, int32_t *);
extern int64_t glXSwapBuffersMscOML(Display *, GLXDrawable, int64_t, int64_t, int64_t);
extern Bool glXWaitForMscOML(Display *, GLXDrawable, int64_t, int64_t, int64_t, int64_t *, int64_t *, int64_t *);
extern Bool glXWaitForSbcOML(Display *, GLXDrawable, int64_t, int64_t *, int64_t *, int64_t *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef Bool(* PFNGLXGETSYNCVALUESOMLPROC)(Display *dpy, GLXDrawable drawable, int64_t *ust, int64_t *msc, int64_t *sbc);
typedef Bool(* PFNGLXGETMSCRATEOMLPROC)(Display *dpy, GLXDrawable drawable, int32_t *numerator, int32_t *denominator);
typedef int64_t (* PFNGLXSWAPBUFFERSMSCOMLPROC)(Display *dpy, GLXDrawable drawable, int64_t target_msc, int64_t divisor, int64_t remainder);
typedef Bool(* PFNGLXWAITFORMSCOMLPROC)(Display *dpy, GLXDrawable drawable, int64_t target_msc, int64_t divisor, int64_t remainder, int64_t *ust, int64_t *msc, int64_t *sbc);
typedef Bool(* PFNGLXWAITFORSBCOMLPROC)(Display *dpy, GLXDrawable drawable, int64_t target_sbc, int64_t *ust, int64_t *msc, int64_t *sbc);
#endif
#ifndef GLX_NV_float_buffer
#define GLX_NV_float_buffer 1
#endif
#ifndef GLX_SGIX_hyperpipe
#define GLX_SGIX_hyperpipe 1
typedef struct
{
char pipeName[GLX_HYPERPIPE_PIPE_NAME_LENGTH_SGIX];
int networkId;
} GLXHyperpipeNetworkSGIX;
typedef struct
{
char pipeName[GLX_HYPERPIPE_PIPE_NAME_LENGTH_SGIX];
int channel;
unsigned int
participationType;
int timeSlice;
} GLXHyperpipeConfigSGIX;
typedef struct
{
char pipeName[GLX_HYPERPIPE_PIPE_NAME_LENGTH_SGIX];
int srcXOrigin, srcYOrigin, srcWidth, srcHeight;
int destXOrigin, destYOrigin, destWidth, destHeight;
} GLXPipeRect;
typedef struct
{
char pipeName[GLX_HYPERPIPE_PIPE_NAME_LENGTH_SGIX];
int XOrigin, YOrigin, maxHeight, maxWidth;
} GLXPipeRectLimits;
#ifdef GLX_GLXEXT_PROTOTYPES
extern GLXHyperpipeNetworkSGIX *glXQueryHyperpipeNetworkSGIX(Display *, int *);
extern int glXHyperpipeConfigSGIX(Display *, int, int, GLXHyperpipeConfigSGIX *, int *);
extern GLXHyperpipeConfigSGIX *glXQueryHyperpipeConfigSGIX(Display *, int, int *);
extern int glXDestroyHyperpipeConfigSGIX(Display *, int);
extern int glXBindHyperpipeSGIX(Display *, int);
extern int glXQueryHyperpipeBestAttribSGIX(Display *, int, int, int, void *, void *);
extern int glXHyperpipeAttribSGIX(Display *, int, int, int, void *);
extern int glXQueryHyperpipeAttribSGIX(Display *, int, int, int, void *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef GLXHyperpipeNetworkSGIX *(* PFNGLXQUERYHYPERPIPENETWORKSGIXPROC)(Display *dpy, int *npipes);
typedef int (* PFNGLXHYPERPIPECONFIGSGIXPROC)(Display *dpy, int networkId, int npipes, GLXHyperpipeConfigSGIX *cfg, int *hpId);
typedef GLXHyperpipeConfigSGIX *(* PFNGLXQUERYHYPERPIPECONFIGSGIXPROC)(Display *dpy, int hpId, int *npipes);
typedef int (* PFNGLXDESTROYHYPERPIPECONFIGSGIXPROC)(Display *dpy, int hpId);
typedef int (* PFNGLXBINDHYPERPIPESGIXPROC)(Display *dpy, int hpId);
typedef int (* PFNGLXQUERYHYPERPIPEBESTATTRIBSGIXPROC)(Display *dpy, int timeSlice, int attrib, int size, void *attribList, void *returnAttribList);
typedef int (* PFNGLXHYPERPIPEATTRIBSGIXPROC)(Display *dpy, int timeSlice, int attrib, int size, void *attribList);
typedef int (* PFNGLXQUERYHYPERPIPEATTRIBSGIXPROC)(Display *dpy, int timeSlice, int attrib, int size, void *returnAttribList);
#endif
#ifndef GLX_MESA_agp_offset
#define GLX_MESA_agp_offset 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern unsigned int glXGetAGPOffsetMESA(const void *);
#endif /* GLX_GLXEXT_PROTOTYPES */
typedef unsigned int (* PFNGLXGETAGPOFFSETMESAPROC)(const void *pointer);
#endif
/*
* GLX_NV_vertex_array_range is not a real extension name...
*/
#ifndef GLX_NV_vertex_array_range
#define GLX_NV_vertex_array_range 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern void *glXAllocateMemoryNV(GLsizei size, GLfloat readfreq,
GLfloat writefreq, GLfloat priority);
extern void glXFreeMemoryNV(GLvoid *pointer);
#endif
typedef void *(* PFNGLXALLOCATEMEMORYNVPROC)(GLsizei size,
GLfloat readfreq,
GLfloat writefreq,
GLfloat priority);
typedef void (* PFNGLXFREEMEMORYNVPROC)(GLvoid *pointer);
#endif
#ifndef GLX_NV_swap_group
#define GLX_NV_swap_group 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern Bool glXJoinSwapGroupNV(Display *dpy, GLXDrawable drawable,
GLuint group);
extern Bool glXBindSwapBarrierNV(Display *dpy, GLuint group, GLuint barrier);
extern Bool glXQuerySwapGroupNV(Display *dpy, GLXDrawable drawable,
GLuint *group, GLuint *barrier);
extern Bool glXQueryMaxSwapGroupsNV(Display *dpy, int screen,
GLuint *maxGroups, GLuint *maxBarriers);
extern Bool glXQueryFrameCountNV(Display *dpy, int screen, GLuint *count);
extern Bool glXResetFrameCountNV(Display *dpy, int screen);
#endif
typedef Bool(* PFNGLXJOINSWAPGROUPNVPROC)(Display *dpy,
GLXDrawable drawable,
GLuint group);
typedef Bool(* PFNGLXBINDSWAPBARRIERNVPROC)(Display *dpy,
GLuint group,
GLuint barrier);
typedef Bool(* PFNGLXQUERYSWAPGROUPNVPROC)(Display *dpy,
GLXDrawable drawable,
GLuint *group,
GLuint *barrier);
typedef Bool(* PFNGLXQUERYMAXSWAPGROUPSNVPROC)(Display *dpy,
int screen,
GLuint *maxGroups,
GLuint *maxBarriers);
typedef Bool(* PFNGLXQUERYFRAMECOUNTNVPROC)(Display *dpy,
int screen,
GLuint *count);
typedef Bool(* PFNGLXRESETFRAMECOUNTNVPROC)(Display *dpy, int screen);
#endif
#ifndef GLX_NV_video_out
#define GLX_NV_video_out 1
#ifdef GLX_GLXEXT_PROTOTYPES
extern int glXGetVideoDeviceNV(Display *dpy, int screen, int numVideoDevices,
GLXVideoDeviceNV *pVideoDevice);
extern int glXReleaseVideoDeviceNV(Display *dpy, int screen,
GLXVideoDeviceNV VideoDevice);
extern int glXBindVideoImageNV(Display *dpy, GLXVideoDeviceNV VideoDevice,
GLXPbuffer pbuf, int iVideoBuffer);
extern int glXReleaseVideoImageNV(Display *dpy, GLXPbuffer pbuf);
extern int glXSendPbufferToVideoNV(Display *dpy, GLXPbuffer pbuf,
int iBufferType,
unsigned long *pulCounterPbuffer,
GLboolean bBlock);
extern int glXGetVideoInfoNV(Display *dpy, int screen,
GLXVideoDeviceNV VideoDevice,
unsigned long *pulCounterOutputVideo,
unsigned long *pulCounterOutputPbuffer);
#endif
typedef int (* PFNGLXGETVIDEODEVICENVPROC)(Display *dpy,
int screen,
int numVideoDevices,
GLXVideoDeviceNV *pVideoDevice);
typedef int (* PFNGLXRELEASEVIDEODEVICENVPROC)(Display *dpy,
int screen,
GLXVideoDeviceNV VideoDevice);
typedef int (* PFNGLXBINDVIDEOIMAGENVPROC)(Display *dpy,
GLXVideoDeviceNV VideoDevice,
GLXPbuffer pbuf,
int iVideoBuffer);
typedef int (* PFNGLXRELEASEVIDEOIMAGENVPROC)(Display *dpy,
GLXPbuffer pbuf);
typedef int (* PFNGLXSENDPBUFFERTOVIDEONVPROC)(Display *dpy,
GLXPbuffer pbuf,
int iBufferType,
unsigned long *pulCounterPbuffer,
GLboolean bBlock);
typedef int (* PFNGLXGETVIDEOINFONVPROC)(Display *dpy, int screen,
GLXVideoDeviceNV VideoDevice,
unsigned long *pulCounterOutputVideo,
unsigned long *pulCounterOutputPbuffer);
#endif
#ifndef GLX_EXT_texture_from_pixmap
#define GLX_EXT_texture_from_pixmap
#ifdef GLX_GLXEXT_PROTOTYPES
extern void glXBindTexImageEXT(Display *dpy, GLXDrawable drawable,
int buffer, const int *attrib_list);
extern void glXReleaseTexImageEXT(Display *dpy, GLXDrawable drawable,
int buffer);
#endif
typedef void (* PFNGLXBINDTEXIMAGEEXTPROC)(Display *dpy,
GLXDrawable drawable,
int buffer,
const int *attrib_list);
typedef void (* PFNGLXRELEASETEXIMAGEEXTPROC)(Display *dpy,
GLXDrawable drawable,
int buffer);
#endif
#ifdef __cplusplus
}
#endif
#endif

958
Common/GL/wglew.h Normal file
View File

@@ -0,0 +1,958 @@
/*
** The OpenGL Extension Wrangler Library
** Copyright (C) 2002-2006, Milan Ikits <milan ikits[]ieee org>
** Copyright (C) 2002-2006, Marcelo E. Magallon <mmagallo[]debian org>
** Copyright (C) 2002, Lev Povalahev
** All rights reserved.
**
** Redistribution and use in source and binary forms, with or without
** modification, are permitted provided that the following conditions are met:
**
** * Redistributions of source code must retain the above copyright notice,
** this list of conditions and the following disclaimer.
** * Redistributions in binary form must reproduce the above copyright notice,
** this list of conditions and the following disclaimer in the documentation
** and/or other materials provided with the distribution.
** * The name of the author may be used to endorse or promote products
** derived from this software without specific prior written permission.
**
** THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
** AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
** IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
** ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
** LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
** CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
** SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
** INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
** CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
** ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
** THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __wglew_h__
#define __wglew_h__
#define __WGLEW_H__
#ifdef __wglext_h_
#error wglext.h included before wglew.h
#endif
#define __wglext_h_
#if !defined(APIENTRY) && !defined(__CYGWIN__)
# ifndef WIN32_LEAN_AND_MEAN
# define WIN32_LEAN_AND_MEAN 1
# endif
#include <windows.h>
#endif
/*
* GLEW_STATIC needs to be set when using the static version.
* GLEW_BUILD is set when building the DLL version.
*/
#ifdef GLEW_STATIC
# define GLEWAPI extern
#else
# ifdef GLEW_BUILD
# define GLEWAPI extern __declspec(dllexport)
# else
# define GLEWAPI extern __declspec(dllimport)
# endif
#endif
#ifdef __cplusplus
extern "C" {
#endif
/* -------------------------- WGL_3DFX_multisample ------------------------- */
#ifndef WGL_3DFX_multisample
#define WGL_3DFX_multisample 1
#define WGL_SAMPLE_BUFFERS_3DFX 0x2060
#define WGL_SAMPLES_3DFX 0x2061
#define WGLEW_3DFX_multisample WGLEW_GET_VAR(__WGLEW_3DFX_multisample)
#endif /* WGL_3DFX_multisample */
/* ------------------------- WGL_3DL_stereo_control ------------------------ */
#ifndef WGL_3DL_stereo_control
#define WGL_3DL_stereo_control 1
#define WGL_STEREO_EMITTER_ENABLE_3DL 0x2055
#define WGL_STEREO_EMITTER_DISABLE_3DL 0x2056
#define WGL_STEREO_POLARITY_NORMAL_3DL 0x2057
#define WGL_STEREO_POLARITY_INVERT_3DL 0x2058
typedef BOOL (WINAPI *PFNWGLSETSTEREOEMITTERSTATE3DLPROC)(HDC hDC, UINT uState);
#define wglSetStereoEmitterState3DL WGLEW_GET_FUN(__wglewSetStereoEmitterState3DL)
#define WGLEW_3DL_stereo_control WGLEW_GET_VAR(__WGLEW_3DL_stereo_control)
#endif /* WGL_3DL_stereo_control */
/* ------------------------- WGL_ARB_buffer_region ------------------------- */
#ifndef WGL_ARB_buffer_region
#define WGL_ARB_buffer_region 1
#define WGL_FRONT_COLOR_BUFFER_BIT_ARB 0x00000001
#define WGL_BACK_COLOR_BUFFER_BIT_ARB 0x00000002
#define WGL_DEPTH_BUFFER_BIT_ARB 0x00000004
#define WGL_STENCIL_BUFFER_BIT_ARB 0x00000008
typedef HANDLE(WINAPI *PFNWGLCREATEBUFFERREGIONARBPROC)(HDC hDC, int iLayerPlane, UINT uType);
typedef VOID (WINAPI *PFNWGLDELETEBUFFERREGIONARBPROC)(HANDLE hRegion);
typedef BOOL (WINAPI *PFNWGLRESTOREBUFFERREGIONARBPROC)(HANDLE hRegion, int x, int y, int width, int height, int xSrc, int ySrc);
typedef BOOL (WINAPI *PFNWGLSAVEBUFFERREGIONARBPROC)(HANDLE hRegion, int x, int y, int width, int height);
#define wglCreateBufferRegionARB WGLEW_GET_FUN(__wglewCreateBufferRegionARB)
#define wglDeleteBufferRegionARB WGLEW_GET_FUN(__wglewDeleteBufferRegionARB)
#define wglRestoreBufferRegionARB WGLEW_GET_FUN(__wglewRestoreBufferRegionARB)
#define wglSaveBufferRegionARB WGLEW_GET_FUN(__wglewSaveBufferRegionARB)
#define WGLEW_ARB_buffer_region WGLEW_GET_VAR(__WGLEW_ARB_buffer_region)
#endif /* WGL_ARB_buffer_region */
/* ----------------------- WGL_ARB_extensions_string ----------------------- */
#ifndef WGL_ARB_extensions_string
#define WGL_ARB_extensions_string 1
typedef const char *(WINAPI *PFNWGLGETEXTENSIONSSTRINGARBPROC)(HDC hdc);
#define wglGetExtensionsStringARB WGLEW_GET_FUN(__wglewGetExtensionsStringARB)
#define WGLEW_ARB_extensions_string WGLEW_GET_VAR(__WGLEW_ARB_extensions_string)
#endif /* WGL_ARB_extensions_string */
/* ----------------------- WGL_ARB_make_current_read ----------------------- */
#ifndef WGL_ARB_make_current_read
#define WGL_ARB_make_current_read 1
typedef HDC(WINAPI *PFNWGLGETCURRENTREADDCARBPROC)(VOID);
typedef BOOL (WINAPI *PFNWGLMAKECONTEXTCURRENTARBPROC)(HDC hDrawDC, HDC hReadDC, HGLRC hglrc);
#define wglGetCurrentReadDCARB WGLEW_GET_FUN(__wglewGetCurrentReadDCARB)
#define wglMakeContextCurrentARB WGLEW_GET_FUN(__wglewMakeContextCurrentARB)
#define WGLEW_ARB_make_current_read WGLEW_GET_VAR(__WGLEW_ARB_make_current_read)
#endif /* WGL_ARB_make_current_read */
/* -------------------------- WGL_ARB_multisample -------------------------- */
#ifndef WGL_ARB_multisample
#define WGL_ARB_multisample 1
#define WGL_SAMPLE_BUFFERS_ARB 0x2041
#define WGL_SAMPLES_ARB 0x2042
#define WGLEW_ARB_multisample WGLEW_GET_VAR(__WGLEW_ARB_multisample)
#endif /* WGL_ARB_multisample */
/* ---------------------------- WGL_ARB_pbuffer ---------------------------- */
#ifndef WGL_ARB_pbuffer
#define WGL_ARB_pbuffer 1
#define WGL_DRAW_TO_PBUFFER_ARB 0x202D
#define WGL_MAX_PBUFFER_PIXELS_ARB 0x202E
#define WGL_MAX_PBUFFER_WIDTH_ARB 0x202F
#define WGL_MAX_PBUFFER_HEIGHT_ARB 0x2030
#define WGL_PBUFFER_LARGEST_ARB 0x2033
#define WGL_PBUFFER_WIDTH_ARB 0x2034
#define WGL_PBUFFER_HEIGHT_ARB 0x2035
#define WGL_PBUFFER_LOST_ARB 0x2036
DECLARE_HANDLE(HPBUFFERARB);
typedef HPBUFFERARB(WINAPI *PFNWGLCREATEPBUFFERARBPROC)(HDC hDC, int iPixelFormat, int iWidth, int iHeight, const int *piAttribList);
typedef BOOL (WINAPI *PFNWGLDESTROYPBUFFERARBPROC)(HPBUFFERARB hPbuffer);
typedef HDC(WINAPI *PFNWGLGETPBUFFERDCARBPROC)(HPBUFFERARB hPbuffer);
typedef BOOL (WINAPI *PFNWGLQUERYPBUFFERARBPROC)(HPBUFFERARB hPbuffer, int iAttribute, int *piValue);
typedef int (WINAPI *PFNWGLRELEASEPBUFFERDCARBPROC)(HPBUFFERARB hPbuffer, HDC hDC);
#define wglCreatePbufferARB WGLEW_GET_FUN(__wglewCreatePbufferARB)
#define wglDestroyPbufferARB WGLEW_GET_FUN(__wglewDestroyPbufferARB)
#define wglGetPbufferDCARB WGLEW_GET_FUN(__wglewGetPbufferDCARB)
#define wglQueryPbufferARB WGLEW_GET_FUN(__wglewQueryPbufferARB)
#define wglReleasePbufferDCARB WGLEW_GET_FUN(__wglewReleasePbufferDCARB)
#define WGLEW_ARB_pbuffer WGLEW_GET_VAR(__WGLEW_ARB_pbuffer)
#endif /* WGL_ARB_pbuffer */
/* -------------------------- WGL_ARB_pixel_format ------------------------- */
#ifndef WGL_ARB_pixel_format
#define WGL_ARB_pixel_format 1
#define WGL_NUMBER_PIXEL_FORMATS_ARB 0x2000
#define WGL_DRAW_TO_WINDOW_ARB 0x2001
#define WGL_DRAW_TO_BITMAP_ARB 0x2002
#define WGL_ACCELERATION_ARB 0x2003
#define WGL_NEED_PALETTE_ARB 0x2004
#define WGL_NEED_SYSTEM_PALETTE_ARB 0x2005
#define WGL_SWAP_LAYER_BUFFERS_ARB 0x2006
#define WGL_SWAP_METHOD_ARB 0x2007
#define WGL_NUMBER_OVERLAYS_ARB 0x2008
#define WGL_NUMBER_UNDERLAYS_ARB 0x2009
#define WGL_TRANSPARENT_ARB 0x200A
#define WGL_SHARE_DEPTH_ARB 0x200C
#define WGL_SHARE_STENCIL_ARB 0x200D
#define WGL_SHARE_ACCUM_ARB 0x200E
#define WGL_SUPPORT_GDI_ARB 0x200F
#define WGL_SUPPORT_OPENGL_ARB 0x2010
#define WGL_DOUBLE_BUFFER_ARB 0x2011
#define WGL_STEREO_ARB 0x2012
#define WGL_PIXEL_TYPE_ARB 0x2013
#define WGL_COLOR_BITS_ARB 0x2014
#define WGL_RED_BITS_ARB 0x2015
#define WGL_RED_SHIFT_ARB 0x2016
#define WGL_GREEN_BITS_ARB 0x2017
#define WGL_GREEN_SHIFT_ARB 0x2018
#define WGL_BLUE_BITS_ARB 0x2019
#define WGL_BLUE_SHIFT_ARB 0x201A
#define WGL_ALPHA_BITS_ARB 0x201B
#define WGL_ALPHA_SHIFT_ARB 0x201C
#define WGL_ACCUM_BITS_ARB 0x201D
#define WGL_ACCUM_RED_BITS_ARB 0x201E
#define WGL_ACCUM_GREEN_BITS_ARB 0x201F
#define WGL_ACCUM_BLUE_BITS_ARB 0x2020
#define WGL_ACCUM_ALPHA_BITS_ARB 0x2021
#define WGL_DEPTH_BITS_ARB 0x2022
#define WGL_STENCIL_BITS_ARB 0x2023
#define WGL_AUX_BUFFERS_ARB 0x2024
#define WGL_NO_ACCELERATION_ARB 0x2025
#define WGL_GENERIC_ACCELERATION_ARB 0x2026
#define WGL_FULL_ACCELERATION_ARB 0x2027
#define WGL_SWAP_EXCHANGE_ARB 0x2028
#define WGL_SWAP_COPY_ARB 0x2029
#define WGL_SWAP_UNDEFINED_ARB 0x202A
#define WGL_TYPE_RGBA_ARB 0x202B
#define WGL_TYPE_COLORINDEX_ARB 0x202C
#define WGL_TRANSPARENT_RED_VALUE_ARB 0x2037
#define WGL_TRANSPARENT_GREEN_VALUE_ARB 0x2038
#define WGL_TRANSPARENT_BLUE_VALUE_ARB 0x2039
#define WGL_TRANSPARENT_ALPHA_VALUE_ARB 0x203A
#define WGL_TRANSPARENT_INDEX_VALUE_ARB 0x203B
typedef BOOL (WINAPI *PFNWGLCHOOSEPIXELFORMATARBPROC)(HDC hdc, const int *piAttribIList, const FLOAT *pfAttribFList, UINT nMaxFormats, int *piFormats, UINT *nNumFormats);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBFVARBPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, const int *piAttributes, FLOAT *pfValues);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBIVARBPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, const int *piAttributes, int *piValues);
#define wglChoosePixelFormatARB WGLEW_GET_FUN(__wglewChoosePixelFormatARB)
#define wglGetPixelFormatAttribfvARB WGLEW_GET_FUN(__wglewGetPixelFormatAttribfvARB)
#define wglGetPixelFormatAttribivARB WGLEW_GET_FUN(__wglewGetPixelFormatAttribivARB)
#define WGLEW_ARB_pixel_format WGLEW_GET_VAR(__WGLEW_ARB_pixel_format)
#endif /* WGL_ARB_pixel_format */
/* ----------------------- WGL_ARB_pixel_format_float ---------------------- */
#ifndef WGL_ARB_pixel_format_float
#define WGL_ARB_pixel_format_float 1
#define WGL_TYPE_RGBA_FLOAT_ARB 0x21A0
#define WGLEW_ARB_pixel_format_float WGLEW_GET_VAR(__WGLEW_ARB_pixel_format_float)
#endif /* WGL_ARB_pixel_format_float */
/* ------------------------- WGL_ARB_render_texture ------------------------ */
#ifndef WGL_ARB_render_texture
#define WGL_ARB_render_texture 1
#define WGL_BIND_TO_TEXTURE_RGB_ARB 0x2070
#define WGL_BIND_TO_TEXTURE_RGBA_ARB 0x2071
#define WGL_TEXTURE_FORMAT_ARB 0x2072
#define WGL_TEXTURE_TARGET_ARB 0x2073
#define WGL_MIPMAP_TEXTURE_ARB 0x2074
#define WGL_TEXTURE_RGB_ARB 0x2075
#define WGL_TEXTURE_RGBA_ARB 0x2076
#define WGL_NO_TEXTURE_ARB 0x2077
#define WGL_TEXTURE_CUBE_MAP_ARB 0x2078
#define WGL_TEXTURE_1D_ARB 0x2079
#define WGL_TEXTURE_2D_ARB 0x207A
#define WGL_MIPMAP_LEVEL_ARB 0x207B
#define WGL_CUBE_MAP_FACE_ARB 0x207C
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_X_ARB 0x207D
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_X_ARB 0x207E
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_Y_ARB 0x207F
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_Y_ARB 0x2080
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_Z_ARB 0x2081
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_Z_ARB 0x2082
#define WGL_FRONT_LEFT_ARB 0x2083
#define WGL_FRONT_RIGHT_ARB 0x2084
#define WGL_BACK_LEFT_ARB 0x2085
#define WGL_BACK_RIGHT_ARB 0x2086
#define WGL_AUX0_ARB 0x2087
#define WGL_AUX1_ARB 0x2088
#define WGL_AUX2_ARB 0x2089
#define WGL_AUX3_ARB 0x208A
#define WGL_AUX4_ARB 0x208B
#define WGL_AUX5_ARB 0x208C
#define WGL_AUX6_ARB 0x208D
#define WGL_AUX7_ARB 0x208E
#define WGL_AUX8_ARB 0x208F
#define WGL_AUX9_ARB 0x2090
typedef BOOL (WINAPI *PFNWGLBINDTEXIMAGEARBPROC)(HPBUFFERARB hPbuffer, int iBuffer);
typedef BOOL (WINAPI *PFNWGLRELEASETEXIMAGEARBPROC)(HPBUFFERARB hPbuffer, int iBuffer);
typedef BOOL (WINAPI *PFNWGLSETPBUFFERATTRIBARBPROC)(HPBUFFERARB hPbuffer, const int *piAttribList);
#define wglBindTexImageARB WGLEW_GET_FUN(__wglewBindTexImageARB)
#define wglReleaseTexImageARB WGLEW_GET_FUN(__wglewReleaseTexImageARB)
#define wglSetPbufferAttribARB WGLEW_GET_FUN(__wglewSetPbufferAttribARB)
#define WGLEW_ARB_render_texture WGLEW_GET_VAR(__WGLEW_ARB_render_texture)
#endif /* WGL_ARB_render_texture */
/* ----------------------- WGL_ATI_pixel_format_float ---------------------- */
#ifndef WGL_ATI_pixel_format_float
#define WGL_ATI_pixel_format_float 1
#define WGL_TYPE_RGBA_FLOAT_ATI 0x21A0
#define GL_RGBA_FLOAT_MODE_ATI 0x8820
#define GL_COLOR_CLEAR_UNCLAMPED_VALUE_ATI 0x8835
#define WGLEW_ATI_pixel_format_float WGLEW_GET_VAR(__WGLEW_ATI_pixel_format_float)
#endif /* WGL_ATI_pixel_format_float */
/* -------------------- WGL_ATI_render_texture_rectangle ------------------- */
#ifndef WGL_ATI_render_texture_rectangle
#define WGL_ATI_render_texture_rectangle 1
#define WGL_TEXTURE_RECTANGLE_ATI 0x21A5
#define WGLEW_ATI_render_texture_rectangle WGLEW_GET_VAR(__WGLEW_ATI_render_texture_rectangle)
#endif /* WGL_ATI_render_texture_rectangle */
/* -------------------------- WGL_EXT_depth_float -------------------------- */
#ifndef WGL_EXT_depth_float
#define WGL_EXT_depth_float 1
#define WGL_DEPTH_FLOAT_EXT 0x2040
#define WGLEW_EXT_depth_float WGLEW_GET_VAR(__WGLEW_EXT_depth_float)
#endif /* WGL_EXT_depth_float */
/* ---------------------- WGL_EXT_display_color_table ---------------------- */
#ifndef WGL_EXT_display_color_table
#define WGL_EXT_display_color_table 1
typedef GLboolean(WINAPI *PFNWGLBINDDISPLAYCOLORTABLEEXTPROC)(GLushort id);
typedef GLboolean(WINAPI *PFNWGLCREATEDISPLAYCOLORTABLEEXTPROC)(GLushort id);
typedef void (WINAPI *PFNWGLDESTROYDISPLAYCOLORTABLEEXTPROC)(GLushort id);
typedef GLboolean(WINAPI *PFNWGLLOADDISPLAYCOLORTABLEEXTPROC)(GLushort *table, GLuint length);
#define wglBindDisplayColorTableEXT WGLEW_GET_FUN(__wglewBindDisplayColorTableEXT)
#define wglCreateDisplayColorTableEXT WGLEW_GET_FUN(__wglewCreateDisplayColorTableEXT)
#define wglDestroyDisplayColorTableEXT WGLEW_GET_FUN(__wglewDestroyDisplayColorTableEXT)
#define wglLoadDisplayColorTableEXT WGLEW_GET_FUN(__wglewLoadDisplayColorTableEXT)
#define WGLEW_EXT_display_color_table WGLEW_GET_VAR(__WGLEW_EXT_display_color_table)
#endif /* WGL_EXT_display_color_table */
/* ----------------------- WGL_EXT_extensions_string ----------------------- */
#ifndef WGL_EXT_extensions_string
#define WGL_EXT_extensions_string 1
typedef const char *(WINAPI *PFNWGLGETEXTENSIONSSTRINGEXTPROC)(void);
#define wglGetExtensionsStringEXT WGLEW_GET_FUN(__wglewGetExtensionsStringEXT)
#define WGLEW_EXT_extensions_string WGLEW_GET_VAR(__WGLEW_EXT_extensions_string)
#endif /* WGL_EXT_extensions_string */
/* ------------------------ WGL_EXT_framebuffer_sRGB ----------------------- */
#ifndef WGL_EXT_framebuffer_sRGB
#define WGL_EXT_framebuffer_sRGB 1
#define WGL_FRAMEBUFFER_SRGB_CAPABLE_EXT 0x20A9
#define WGLEW_EXT_framebuffer_sRGB WGLEW_GET_VAR(__WGLEW_EXT_framebuffer_sRGB)
#endif /* WGL_EXT_framebuffer_sRGB */
/* ----------------------- WGL_EXT_make_current_read ----------------------- */
#ifndef WGL_EXT_make_current_read
#define WGL_EXT_make_current_read 1
typedef HDC(WINAPI *PFNWGLGETCURRENTREADDCEXTPROC)(VOID);
typedef BOOL (WINAPI *PFNWGLMAKECONTEXTCURRENTEXTPROC)(HDC hDrawDC, HDC hReadDC, HGLRC hglrc);
#define wglGetCurrentReadDCEXT WGLEW_GET_FUN(__wglewGetCurrentReadDCEXT)
#define wglMakeContextCurrentEXT WGLEW_GET_FUN(__wglewMakeContextCurrentEXT)
#define WGLEW_EXT_make_current_read WGLEW_GET_VAR(__WGLEW_EXT_make_current_read)
#endif /* WGL_EXT_make_current_read */
/* -------------------------- WGL_EXT_multisample -------------------------- */
#ifndef WGL_EXT_multisample
#define WGL_EXT_multisample 1
#define WGL_SAMPLE_BUFFERS_EXT 0x2041
#define WGL_SAMPLES_EXT 0x2042
#define WGLEW_EXT_multisample WGLEW_GET_VAR(__WGLEW_EXT_multisample)
#endif /* WGL_EXT_multisample */
/* ---------------------------- WGL_EXT_pbuffer ---------------------------- */
#ifndef WGL_EXT_pbuffer
#define WGL_EXT_pbuffer 1
#define WGL_DRAW_TO_PBUFFER_EXT 0x202D
#define WGL_MAX_PBUFFER_PIXELS_EXT 0x202E
#define WGL_MAX_PBUFFER_WIDTH_EXT 0x202F
#define WGL_MAX_PBUFFER_HEIGHT_EXT 0x2030
#define WGL_OPTIMAL_PBUFFER_WIDTH_EXT 0x2031
#define WGL_OPTIMAL_PBUFFER_HEIGHT_EXT 0x2032
#define WGL_PBUFFER_LARGEST_EXT 0x2033
#define WGL_PBUFFER_WIDTH_EXT 0x2034
#define WGL_PBUFFER_HEIGHT_EXT 0x2035
DECLARE_HANDLE(HPBUFFEREXT);
typedef HPBUFFEREXT(WINAPI *PFNWGLCREATEPBUFFEREXTPROC)(HDC hDC, int iPixelFormat, int iWidth, int iHeight, const int *piAttribList);
typedef BOOL (WINAPI *PFNWGLDESTROYPBUFFEREXTPROC)(HPBUFFEREXT hPbuffer);
typedef HDC(WINAPI *PFNWGLGETPBUFFERDCEXTPROC)(HPBUFFEREXT hPbuffer);
typedef BOOL (WINAPI *PFNWGLQUERYPBUFFEREXTPROC)(HPBUFFEREXT hPbuffer, int iAttribute, int *piValue);
typedef int (WINAPI *PFNWGLRELEASEPBUFFERDCEXTPROC)(HPBUFFEREXT hPbuffer, HDC hDC);
#define wglCreatePbufferEXT WGLEW_GET_FUN(__wglewCreatePbufferEXT)
#define wglDestroyPbufferEXT WGLEW_GET_FUN(__wglewDestroyPbufferEXT)
#define wglGetPbufferDCEXT WGLEW_GET_FUN(__wglewGetPbufferDCEXT)
#define wglQueryPbufferEXT WGLEW_GET_FUN(__wglewQueryPbufferEXT)
#define wglReleasePbufferDCEXT WGLEW_GET_FUN(__wglewReleasePbufferDCEXT)
#define WGLEW_EXT_pbuffer WGLEW_GET_VAR(__WGLEW_EXT_pbuffer)
#endif /* WGL_EXT_pbuffer */
/* -------------------------- WGL_EXT_pixel_format ------------------------- */
#ifndef WGL_EXT_pixel_format
#define WGL_EXT_pixel_format 1
#define WGL_NUMBER_PIXEL_FORMATS_EXT 0x2000
#define WGL_DRAW_TO_WINDOW_EXT 0x2001
#define WGL_DRAW_TO_BITMAP_EXT 0x2002
#define WGL_ACCELERATION_EXT 0x2003
#define WGL_NEED_PALETTE_EXT 0x2004
#define WGL_NEED_SYSTEM_PALETTE_EXT 0x2005
#define WGL_SWAP_LAYER_BUFFERS_EXT 0x2006
#define WGL_SWAP_METHOD_EXT 0x2007
#define WGL_NUMBER_OVERLAYS_EXT 0x2008
#define WGL_NUMBER_UNDERLAYS_EXT 0x2009
#define WGL_TRANSPARENT_EXT 0x200A
#define WGL_TRANSPARENT_VALUE_EXT 0x200B
#define WGL_SHARE_DEPTH_EXT 0x200C
#define WGL_SHARE_STENCIL_EXT 0x200D
#define WGL_SHARE_ACCUM_EXT 0x200E
#define WGL_SUPPORT_GDI_EXT 0x200F
#define WGL_SUPPORT_OPENGL_EXT 0x2010
#define WGL_DOUBLE_BUFFER_EXT 0x2011
#define WGL_STEREO_EXT 0x2012
#define WGL_PIXEL_TYPE_EXT 0x2013
#define WGL_COLOR_BITS_EXT 0x2014
#define WGL_RED_BITS_EXT 0x2015
#define WGL_RED_SHIFT_EXT 0x2016
#define WGL_GREEN_BITS_EXT 0x2017
#define WGL_GREEN_SHIFT_EXT 0x2018
#define WGL_BLUE_BITS_EXT 0x2019
#define WGL_BLUE_SHIFT_EXT 0x201A
#define WGL_ALPHA_BITS_EXT 0x201B
#define WGL_ALPHA_SHIFT_EXT 0x201C
#define WGL_ACCUM_BITS_EXT 0x201D
#define WGL_ACCUM_RED_BITS_EXT 0x201E
#define WGL_ACCUM_GREEN_BITS_EXT 0x201F
#define WGL_ACCUM_BLUE_BITS_EXT 0x2020
#define WGL_ACCUM_ALPHA_BITS_EXT 0x2021
#define WGL_DEPTH_BITS_EXT 0x2022
#define WGL_STENCIL_BITS_EXT 0x2023
#define WGL_AUX_BUFFERS_EXT 0x2024
#define WGL_NO_ACCELERATION_EXT 0x2025
#define WGL_GENERIC_ACCELERATION_EXT 0x2026
#define WGL_FULL_ACCELERATION_EXT 0x2027
#define WGL_SWAP_EXCHANGE_EXT 0x2028
#define WGL_SWAP_COPY_EXT 0x2029
#define WGL_SWAP_UNDEFINED_EXT 0x202A
#define WGL_TYPE_RGBA_EXT 0x202B
#define WGL_TYPE_COLORINDEX_EXT 0x202C
typedef BOOL (WINAPI *PFNWGLCHOOSEPIXELFORMATEXTPROC)(HDC hdc, const int *piAttribIList, const FLOAT *pfAttribFList, UINT nMaxFormats, int *piFormats, UINT *nNumFormats);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBFVEXTPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, int *piAttributes, FLOAT *pfValues);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBIVEXTPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, int *piAttributes, int *piValues);
#define wglChoosePixelFormatEXT WGLEW_GET_FUN(__wglewChoosePixelFormatEXT)
#define wglGetPixelFormatAttribfvEXT WGLEW_GET_FUN(__wglewGetPixelFormatAttribfvEXT)
#define wglGetPixelFormatAttribivEXT WGLEW_GET_FUN(__wglewGetPixelFormatAttribivEXT)
#define WGLEW_EXT_pixel_format WGLEW_GET_VAR(__WGLEW_EXT_pixel_format)
#endif /* WGL_EXT_pixel_format */
/* ------------------- WGL_EXT_pixel_format_packed_float ------------------- */
#ifndef WGL_EXT_pixel_format_packed_float
#define WGL_EXT_pixel_format_packed_float 1
#define WGL_TYPE_RGBA_UNSIGNED_FLOAT_EXT 0x20A8
#define WGLEW_EXT_pixel_format_packed_float WGLEW_GET_VAR(__WGLEW_EXT_pixel_format_packed_float)
#endif /* WGL_EXT_pixel_format_packed_float */
/* -------------------------- WGL_EXT_swap_control ------------------------- */
#ifndef WGL_EXT_swap_control
#define WGL_EXT_swap_control 1
typedef int (WINAPI *PFNWGLGETSWAPINTERVALEXTPROC)(void);
typedef BOOL (WINAPI *PFNWGLSWAPINTERVALEXTPROC)(int interval);
#define wglGetSwapIntervalEXT WGLEW_GET_FUN(__wglewGetSwapIntervalEXT)
#define wglSwapIntervalEXT WGLEW_GET_FUN(__wglewSwapIntervalEXT)
#define WGLEW_EXT_swap_control WGLEW_GET_VAR(__WGLEW_EXT_swap_control)
#endif /* WGL_EXT_swap_control */
/* --------------------- WGL_I3D_digital_video_control --------------------- */
#ifndef WGL_I3D_digital_video_control
#define WGL_I3D_digital_video_control 1
#define WGL_DIGITAL_VIDEO_CURSOR_ALPHA_FRAMEBUFFER_I3D 0x2050
#define WGL_DIGITAL_VIDEO_CURSOR_ALPHA_VALUE_I3D 0x2051
#define WGL_DIGITAL_VIDEO_CURSOR_INCLUDED_I3D 0x2052
#define WGL_DIGITAL_VIDEO_GAMMA_CORRECTED_I3D 0x2053
typedef BOOL (WINAPI *PFNWGLGETDIGITALVIDEOPARAMETERSI3DPROC)(HDC hDC, int iAttribute, int *piValue);
typedef BOOL (WINAPI *PFNWGLSETDIGITALVIDEOPARAMETERSI3DPROC)(HDC hDC, int iAttribute, const int *piValue);
#define wglGetDigitalVideoParametersI3D WGLEW_GET_FUN(__wglewGetDigitalVideoParametersI3D)
#define wglSetDigitalVideoParametersI3D WGLEW_GET_FUN(__wglewSetDigitalVideoParametersI3D)
#define WGLEW_I3D_digital_video_control WGLEW_GET_VAR(__WGLEW_I3D_digital_video_control)
#endif /* WGL_I3D_digital_video_control */
/* ----------------------------- WGL_I3D_gamma ----------------------------- */
#ifndef WGL_I3D_gamma
#define WGL_I3D_gamma 1
#define WGL_GAMMA_TABLE_SIZE_I3D 0x204E
#define WGL_GAMMA_EXCLUDE_DESKTOP_I3D 0x204F
typedef BOOL (WINAPI *PFNWGLGETGAMMATABLEI3DPROC)(HDC hDC, int iEntries, USHORT *puRed, USHORT *puGreen, USHORT *puBlue);
typedef BOOL (WINAPI *PFNWGLGETGAMMATABLEPARAMETERSI3DPROC)(HDC hDC, int iAttribute, int *piValue);
typedef BOOL (WINAPI *PFNWGLSETGAMMATABLEI3DPROC)(HDC hDC, int iEntries, const USHORT *puRed, const USHORT *puGreen, const USHORT *puBlue);
typedef BOOL (WINAPI *PFNWGLSETGAMMATABLEPARAMETERSI3DPROC)(HDC hDC, int iAttribute, const int *piValue);
#define wglGetGammaTableI3D WGLEW_GET_FUN(__wglewGetGammaTableI3D)
#define wglGetGammaTableParametersI3D WGLEW_GET_FUN(__wglewGetGammaTableParametersI3D)
#define wglSetGammaTableI3D WGLEW_GET_FUN(__wglewSetGammaTableI3D)
#define wglSetGammaTableParametersI3D WGLEW_GET_FUN(__wglewSetGammaTableParametersI3D)
#define WGLEW_I3D_gamma WGLEW_GET_VAR(__WGLEW_I3D_gamma)
#endif /* WGL_I3D_gamma */
/* ---------------------------- WGL_I3D_genlock ---------------------------- */
#ifndef WGL_I3D_genlock
#define WGL_I3D_genlock 1
#define WGL_GENLOCK_SOURCE_MULTIVIEW_I3D 0x2044
#define WGL_GENLOCK_SOURCE_EXTERNAL_SYNC_I3D 0x2045
#define WGL_GENLOCK_SOURCE_EXTERNAL_FIELD_I3D 0x2046
#define WGL_GENLOCK_SOURCE_EXTERNAL_TTL_I3D 0x2047
#define WGL_GENLOCK_SOURCE_DIGITAL_SYNC_I3D 0x2048
#define WGL_GENLOCK_SOURCE_DIGITAL_FIELD_I3D 0x2049
#define WGL_GENLOCK_SOURCE_EDGE_FALLING_I3D 0x204A
#define WGL_GENLOCK_SOURCE_EDGE_RISING_I3D 0x204B
#define WGL_GENLOCK_SOURCE_EDGE_BOTH_I3D 0x204C
typedef BOOL (WINAPI *PFNWGLDISABLEGENLOCKI3DPROC)(HDC hDC);
typedef BOOL (WINAPI *PFNWGLENABLEGENLOCKI3DPROC)(HDC hDC);
typedef BOOL (WINAPI *PFNWGLGENLOCKSAMPLERATEI3DPROC)(HDC hDC, UINT uRate);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEDELAYI3DPROC)(HDC hDC, UINT uDelay);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEEDGEI3DPROC)(HDC hDC, UINT uEdge);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEI3DPROC)(HDC hDC, UINT uSource);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSAMPLERATEI3DPROC)(HDC hDC, UINT *uRate);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEDELAYI3DPROC)(HDC hDC, UINT *uDelay);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEEDGEI3DPROC)(HDC hDC, UINT *uEdge);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEI3DPROC)(HDC hDC, UINT *uSource);
typedef BOOL (WINAPI *PFNWGLISENABLEDGENLOCKI3DPROC)(HDC hDC, BOOL *pFlag);
typedef BOOL (WINAPI *PFNWGLQUERYGENLOCKMAXSOURCEDELAYI3DPROC)(HDC hDC, UINT *uMaxLineDelay, UINT *uMaxPixelDelay);
#define wglDisableGenlockI3D WGLEW_GET_FUN(__wglewDisableGenlockI3D)
#define wglEnableGenlockI3D WGLEW_GET_FUN(__wglewEnableGenlockI3D)
#define wglGenlockSampleRateI3D WGLEW_GET_FUN(__wglewGenlockSampleRateI3D)
#define wglGenlockSourceDelayI3D WGLEW_GET_FUN(__wglewGenlockSourceDelayI3D)
#define wglGenlockSourceEdgeI3D WGLEW_GET_FUN(__wglewGenlockSourceEdgeI3D)
#define wglGenlockSourceI3D WGLEW_GET_FUN(__wglewGenlockSourceI3D)
#define wglGetGenlockSampleRateI3D WGLEW_GET_FUN(__wglewGetGenlockSampleRateI3D)
#define wglGetGenlockSourceDelayI3D WGLEW_GET_FUN(__wglewGetGenlockSourceDelayI3D)
#define wglGetGenlockSourceEdgeI3D WGLEW_GET_FUN(__wglewGetGenlockSourceEdgeI3D)
#define wglGetGenlockSourceI3D WGLEW_GET_FUN(__wglewGetGenlockSourceI3D)
#define wglIsEnabledGenlockI3D WGLEW_GET_FUN(__wglewIsEnabledGenlockI3D)
#define wglQueryGenlockMaxSourceDelayI3D WGLEW_GET_FUN(__wglewQueryGenlockMaxSourceDelayI3D)
#define WGLEW_I3D_genlock WGLEW_GET_VAR(__WGLEW_I3D_genlock)
#endif /* WGL_I3D_genlock */
/* -------------------------- WGL_I3D_image_buffer ------------------------- */
#ifndef WGL_I3D_image_buffer
#define WGL_I3D_image_buffer 1
#define WGL_IMAGE_BUFFER_MIN_ACCESS_I3D 0x00000001
#define WGL_IMAGE_BUFFER_LOCK_I3D 0x00000002
typedef BOOL (WINAPI *PFNWGLASSOCIATEIMAGEBUFFEREVENTSI3DPROC)(HDC hdc, HANDLE *pEvent, LPVOID *pAddress, DWORD *pSize, UINT count);
typedef LPVOID (WINAPI *PFNWGLCREATEIMAGEBUFFERI3DPROC)(HDC hDC, DWORD dwSize, UINT uFlags);
typedef BOOL (WINAPI *PFNWGLDESTROYIMAGEBUFFERI3DPROC)(HDC hDC, LPVOID pAddress);
typedef BOOL (WINAPI *PFNWGLRELEASEIMAGEBUFFEREVENTSI3DPROC)(HDC hdc, LPVOID *pAddress, UINT count);
#define wglAssociateImageBufferEventsI3D WGLEW_GET_FUN(__wglewAssociateImageBufferEventsI3D)
#define wglCreateImageBufferI3D WGLEW_GET_FUN(__wglewCreateImageBufferI3D)
#define wglDestroyImageBufferI3D WGLEW_GET_FUN(__wglewDestroyImageBufferI3D)
#define wglReleaseImageBufferEventsI3D WGLEW_GET_FUN(__wglewReleaseImageBufferEventsI3D)
#define WGLEW_I3D_image_buffer WGLEW_GET_VAR(__WGLEW_I3D_image_buffer)
#endif /* WGL_I3D_image_buffer */
/* ------------------------ WGL_I3D_swap_frame_lock ------------------------ */
#ifndef WGL_I3D_swap_frame_lock
#define WGL_I3D_swap_frame_lock 1
typedef BOOL (WINAPI *PFNWGLDISABLEFRAMELOCKI3DPROC)(VOID);
typedef BOOL (WINAPI *PFNWGLENABLEFRAMELOCKI3DPROC)(VOID);
typedef BOOL (WINAPI *PFNWGLISENABLEDFRAMELOCKI3DPROC)(BOOL *pFlag);
typedef BOOL (WINAPI *PFNWGLQUERYFRAMELOCKMASTERI3DPROC)(BOOL *pFlag);
#define wglDisableFrameLockI3D WGLEW_GET_FUN(__wglewDisableFrameLockI3D)
#define wglEnableFrameLockI3D WGLEW_GET_FUN(__wglewEnableFrameLockI3D)
#define wglIsEnabledFrameLockI3D WGLEW_GET_FUN(__wglewIsEnabledFrameLockI3D)
#define wglQueryFrameLockMasterI3D WGLEW_GET_FUN(__wglewQueryFrameLockMasterI3D)
#define WGLEW_I3D_swap_frame_lock WGLEW_GET_VAR(__WGLEW_I3D_swap_frame_lock)
#endif /* WGL_I3D_swap_frame_lock */
/* ------------------------ WGL_I3D_swap_frame_usage ----------------------- */
#ifndef WGL_I3D_swap_frame_usage
#define WGL_I3D_swap_frame_usage 1
typedef BOOL (WINAPI *PFNWGLBEGINFRAMETRACKINGI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLENDFRAMETRACKINGI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLGETFRAMEUSAGEI3DPROC)(float *pUsage);
typedef BOOL (WINAPI *PFNWGLQUERYFRAMETRACKINGI3DPROC)(DWORD *pFrameCount, DWORD *pMissedFrames, float *pLastMissedUsage);
#define wglBeginFrameTrackingI3D WGLEW_GET_FUN(__wglewBeginFrameTrackingI3D)
#define wglEndFrameTrackingI3D WGLEW_GET_FUN(__wglewEndFrameTrackingI3D)
#define wglGetFrameUsageI3D WGLEW_GET_FUN(__wglewGetFrameUsageI3D)
#define wglQueryFrameTrackingI3D WGLEW_GET_FUN(__wglewQueryFrameTrackingI3D)
#define WGLEW_I3D_swap_frame_usage WGLEW_GET_VAR(__WGLEW_I3D_swap_frame_usage)
#endif /* WGL_I3D_swap_frame_usage */
/* -------------------------- WGL_NV_float_buffer -------------------------- */
#ifndef WGL_NV_float_buffer
#define WGL_NV_float_buffer 1
#define WGL_FLOAT_COMPONENTS_NV 0x20B0
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_R_NV 0x20B1
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RG_NV 0x20B2
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RGB_NV 0x20B3
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RGBA_NV 0x20B4
#define WGL_TEXTURE_FLOAT_R_NV 0x20B5
#define WGL_TEXTURE_FLOAT_RG_NV 0x20B6
#define WGL_TEXTURE_FLOAT_RGB_NV 0x20B7
#define WGL_TEXTURE_FLOAT_RGBA_NV 0x20B8
#define WGLEW_NV_float_buffer WGLEW_GET_VAR(__WGLEW_NV_float_buffer)
#endif /* WGL_NV_float_buffer */
/* ---------------------- WGL_NV_render_depth_texture ---------------------- */
#ifndef WGL_NV_render_depth_texture
#define WGL_NV_render_depth_texture 1
#define WGL_NO_TEXTURE_ARB 0x2077
#define WGL_BIND_TO_TEXTURE_DEPTH_NV 0x20A3
#define WGL_BIND_TO_TEXTURE_RECTANGLE_DEPTH_NV 0x20A4
#define WGL_DEPTH_TEXTURE_FORMAT_NV 0x20A5
#define WGL_TEXTURE_DEPTH_COMPONENT_NV 0x20A6
#define WGL_DEPTH_COMPONENT_NV 0x20A7
#define WGLEW_NV_render_depth_texture WGLEW_GET_VAR(__WGLEW_NV_render_depth_texture)
#endif /* WGL_NV_render_depth_texture */
/* -------------------- WGL_NV_render_texture_rectangle -------------------- */
#ifndef WGL_NV_render_texture_rectangle
#define WGL_NV_render_texture_rectangle 1
#define WGL_BIND_TO_TEXTURE_RECTANGLE_RGB_NV 0x20A0
#define WGL_BIND_TO_TEXTURE_RECTANGLE_RGBA_NV 0x20A1
#define WGL_TEXTURE_RECTANGLE_NV 0x20A2
#define WGLEW_NV_render_texture_rectangle WGLEW_GET_VAR(__WGLEW_NV_render_texture_rectangle)
#endif /* WGL_NV_render_texture_rectangle */
/* ----------------------- WGL_NV_vertex_array_range ----------------------- */
#ifndef WGL_NV_vertex_array_range
#define WGL_NV_vertex_array_range 1
typedef void *(WINAPI *PFNWGLALLOCATEMEMORYNVPROC)(GLsizei size, GLfloat readFrequency, GLfloat writeFrequency, GLfloat priority);
typedef void (WINAPI *PFNWGLFREEMEMORYNVPROC)(void *pointer);
#define wglAllocateMemoryNV WGLEW_GET_FUN(__wglewAllocateMemoryNV)
#define wglFreeMemoryNV WGLEW_GET_FUN(__wglewFreeMemoryNV)
#define WGLEW_NV_vertex_array_range WGLEW_GET_VAR(__WGLEW_NV_vertex_array_range)
#endif /* WGL_NV_vertex_array_range */
/* -------------------------- WGL_OML_sync_control ------------------------- */
#ifndef WGL_OML_sync_control
#define WGL_OML_sync_control 1
typedef BOOL (WINAPI *PFNWGLGETMSCRATEOMLPROC)(HDC hdc, INT32 *numerator, INT32 *denominator);
typedef BOOL (WINAPI *PFNWGLGETSYNCVALUESOMLPROC)(HDC hdc, INT64 *ust, INT64 *msc, INT64 *sbc);
typedef INT64(WINAPI *PFNWGLSWAPBUFFERSMSCOMLPROC)(HDC hdc, INT64 target_msc, INT64 divisor, INT64 remainder);
typedef INT64(WINAPI *PFNWGLSWAPLAYERBUFFERSMSCOMLPROC)(HDC hdc, INT fuPlanes, INT64 target_msc, INT64 divisor, INT64 remainder);
typedef BOOL (WINAPI *PFNWGLWAITFORMSCOMLPROC)(HDC hdc, INT64 target_msc, INT64 divisor, INT64 remainder, INT64 *ust, INT64 *msc, INT64 *sbc);
typedef BOOL (WINAPI *PFNWGLWAITFORSBCOMLPROC)(HDC hdc, INT64 target_sbc, INT64 *ust, INT64 *msc, INT64 *sbc);
#define wglGetMscRateOML WGLEW_GET_FUN(__wglewGetMscRateOML)
#define wglGetSyncValuesOML WGLEW_GET_FUN(__wglewGetSyncValuesOML)
#define wglSwapBuffersMscOML WGLEW_GET_FUN(__wglewSwapBuffersMscOML)
#define wglSwapLayerBuffersMscOML WGLEW_GET_FUN(__wglewSwapLayerBuffersMscOML)
#define wglWaitForMscOML WGLEW_GET_FUN(__wglewWaitForMscOML)
#define wglWaitForSbcOML WGLEW_GET_FUN(__wglewWaitForSbcOML)
#define WGLEW_OML_sync_control WGLEW_GET_VAR(__WGLEW_OML_sync_control)
#endif /* WGL_OML_sync_control */
/* ------------------------------------------------------------------------- */
#ifdef GLEW_MX
#define WGLEW_EXPORT
#else
#define WGLEW_EXPORT GLEWAPI
#endif /* GLEW_MX */
#ifdef GLEW_MX
struct WGLEWContextStruct
{
#endif /* GLEW_MX */
WGLEW_EXPORT PFNWGLSETSTEREOEMITTERSTATE3DLPROC __wglewSetStereoEmitterState3DL;
WGLEW_EXPORT PFNWGLCREATEBUFFERREGIONARBPROC __wglewCreateBufferRegionARB;
WGLEW_EXPORT PFNWGLDELETEBUFFERREGIONARBPROC __wglewDeleteBufferRegionARB;
WGLEW_EXPORT PFNWGLRESTOREBUFFERREGIONARBPROC __wglewRestoreBufferRegionARB;
WGLEW_EXPORT PFNWGLSAVEBUFFERREGIONARBPROC __wglewSaveBufferRegionARB;
WGLEW_EXPORT PFNWGLGETEXTENSIONSSTRINGARBPROC __wglewGetExtensionsStringARB;
WGLEW_EXPORT PFNWGLGETCURRENTREADDCARBPROC __wglewGetCurrentReadDCARB;
WGLEW_EXPORT PFNWGLMAKECONTEXTCURRENTARBPROC __wglewMakeContextCurrentARB;
WGLEW_EXPORT PFNWGLCREATEPBUFFERARBPROC __wglewCreatePbufferARB;
WGLEW_EXPORT PFNWGLDESTROYPBUFFERARBPROC __wglewDestroyPbufferARB;
WGLEW_EXPORT PFNWGLGETPBUFFERDCARBPROC __wglewGetPbufferDCARB;
WGLEW_EXPORT PFNWGLQUERYPBUFFERARBPROC __wglewQueryPbufferARB;
WGLEW_EXPORT PFNWGLRELEASEPBUFFERDCARBPROC __wglewReleasePbufferDCARB;
WGLEW_EXPORT PFNWGLCHOOSEPIXELFORMATARBPROC __wglewChoosePixelFormatARB;
WGLEW_EXPORT PFNWGLGETPIXELFORMATATTRIBFVARBPROC __wglewGetPixelFormatAttribfvARB;
WGLEW_EXPORT PFNWGLGETPIXELFORMATATTRIBIVARBPROC __wglewGetPixelFormatAttribivARB;
WGLEW_EXPORT PFNWGLBINDTEXIMAGEARBPROC __wglewBindTexImageARB;
WGLEW_EXPORT PFNWGLRELEASETEXIMAGEARBPROC __wglewReleaseTexImageARB;
WGLEW_EXPORT PFNWGLSETPBUFFERATTRIBARBPROC __wglewSetPbufferAttribARB;
WGLEW_EXPORT PFNWGLBINDDISPLAYCOLORTABLEEXTPROC __wglewBindDisplayColorTableEXT;
WGLEW_EXPORT PFNWGLCREATEDISPLAYCOLORTABLEEXTPROC __wglewCreateDisplayColorTableEXT;
WGLEW_EXPORT PFNWGLDESTROYDISPLAYCOLORTABLEEXTPROC __wglewDestroyDisplayColorTableEXT;
WGLEW_EXPORT PFNWGLLOADDISPLAYCOLORTABLEEXTPROC __wglewLoadDisplayColorTableEXT;
WGLEW_EXPORT PFNWGLGETEXTENSIONSSTRINGEXTPROC __wglewGetExtensionsStringEXT;
WGLEW_EXPORT PFNWGLGETCURRENTREADDCEXTPROC __wglewGetCurrentReadDCEXT;
WGLEW_EXPORT PFNWGLMAKECONTEXTCURRENTEXTPROC __wglewMakeContextCurrentEXT;
WGLEW_EXPORT PFNWGLCREATEPBUFFEREXTPROC __wglewCreatePbufferEXT;
WGLEW_EXPORT PFNWGLDESTROYPBUFFEREXTPROC __wglewDestroyPbufferEXT;
WGLEW_EXPORT PFNWGLGETPBUFFERDCEXTPROC __wglewGetPbufferDCEXT;
WGLEW_EXPORT PFNWGLQUERYPBUFFEREXTPROC __wglewQueryPbufferEXT;
WGLEW_EXPORT PFNWGLRELEASEPBUFFERDCEXTPROC __wglewReleasePbufferDCEXT;
WGLEW_EXPORT PFNWGLCHOOSEPIXELFORMATEXTPROC __wglewChoosePixelFormatEXT;
WGLEW_EXPORT PFNWGLGETPIXELFORMATATTRIBFVEXTPROC __wglewGetPixelFormatAttribfvEXT;
WGLEW_EXPORT PFNWGLGETPIXELFORMATATTRIBIVEXTPROC __wglewGetPixelFormatAttribivEXT;
WGLEW_EXPORT PFNWGLGETSWAPINTERVALEXTPROC __wglewGetSwapIntervalEXT;
WGLEW_EXPORT PFNWGLSWAPINTERVALEXTPROC __wglewSwapIntervalEXT;
WGLEW_EXPORT PFNWGLGETDIGITALVIDEOPARAMETERSI3DPROC __wglewGetDigitalVideoParametersI3D;
WGLEW_EXPORT PFNWGLSETDIGITALVIDEOPARAMETERSI3DPROC __wglewSetDigitalVideoParametersI3D;
WGLEW_EXPORT PFNWGLGETGAMMATABLEI3DPROC __wglewGetGammaTableI3D;
WGLEW_EXPORT PFNWGLGETGAMMATABLEPARAMETERSI3DPROC __wglewGetGammaTableParametersI3D;
WGLEW_EXPORT PFNWGLSETGAMMATABLEI3DPROC __wglewSetGammaTableI3D;
WGLEW_EXPORT PFNWGLSETGAMMATABLEPARAMETERSI3DPROC __wglewSetGammaTableParametersI3D;
WGLEW_EXPORT PFNWGLDISABLEGENLOCKI3DPROC __wglewDisableGenlockI3D;
WGLEW_EXPORT PFNWGLENABLEGENLOCKI3DPROC __wglewEnableGenlockI3D;
WGLEW_EXPORT PFNWGLGENLOCKSAMPLERATEI3DPROC __wglewGenlockSampleRateI3D;
WGLEW_EXPORT PFNWGLGENLOCKSOURCEDELAYI3DPROC __wglewGenlockSourceDelayI3D;
WGLEW_EXPORT PFNWGLGENLOCKSOURCEEDGEI3DPROC __wglewGenlockSourceEdgeI3D;
WGLEW_EXPORT PFNWGLGENLOCKSOURCEI3DPROC __wglewGenlockSourceI3D;
WGLEW_EXPORT PFNWGLGETGENLOCKSAMPLERATEI3DPROC __wglewGetGenlockSampleRateI3D;
WGLEW_EXPORT PFNWGLGETGENLOCKSOURCEDELAYI3DPROC __wglewGetGenlockSourceDelayI3D;
WGLEW_EXPORT PFNWGLGETGENLOCKSOURCEEDGEI3DPROC __wglewGetGenlockSourceEdgeI3D;
WGLEW_EXPORT PFNWGLGETGENLOCKSOURCEI3DPROC __wglewGetGenlockSourceI3D;
WGLEW_EXPORT PFNWGLISENABLEDGENLOCKI3DPROC __wglewIsEnabledGenlockI3D;
WGLEW_EXPORT PFNWGLQUERYGENLOCKMAXSOURCEDELAYI3DPROC __wglewQueryGenlockMaxSourceDelayI3D;
WGLEW_EXPORT PFNWGLASSOCIATEIMAGEBUFFEREVENTSI3DPROC __wglewAssociateImageBufferEventsI3D;
WGLEW_EXPORT PFNWGLCREATEIMAGEBUFFERI3DPROC __wglewCreateImageBufferI3D;
WGLEW_EXPORT PFNWGLDESTROYIMAGEBUFFERI3DPROC __wglewDestroyImageBufferI3D;
WGLEW_EXPORT PFNWGLRELEASEIMAGEBUFFEREVENTSI3DPROC __wglewReleaseImageBufferEventsI3D;
WGLEW_EXPORT PFNWGLDISABLEFRAMELOCKI3DPROC __wglewDisableFrameLockI3D;
WGLEW_EXPORT PFNWGLENABLEFRAMELOCKI3DPROC __wglewEnableFrameLockI3D;
WGLEW_EXPORT PFNWGLISENABLEDFRAMELOCKI3DPROC __wglewIsEnabledFrameLockI3D;
WGLEW_EXPORT PFNWGLQUERYFRAMELOCKMASTERI3DPROC __wglewQueryFrameLockMasterI3D;
WGLEW_EXPORT PFNWGLBEGINFRAMETRACKINGI3DPROC __wglewBeginFrameTrackingI3D;
WGLEW_EXPORT PFNWGLENDFRAMETRACKINGI3DPROC __wglewEndFrameTrackingI3D;
WGLEW_EXPORT PFNWGLGETFRAMEUSAGEI3DPROC __wglewGetFrameUsageI3D;
WGLEW_EXPORT PFNWGLQUERYFRAMETRACKINGI3DPROC __wglewQueryFrameTrackingI3D;
WGLEW_EXPORT PFNWGLALLOCATEMEMORYNVPROC __wglewAllocateMemoryNV;
WGLEW_EXPORT PFNWGLFREEMEMORYNVPROC __wglewFreeMemoryNV;
WGLEW_EXPORT PFNWGLGETMSCRATEOMLPROC __wglewGetMscRateOML;
WGLEW_EXPORT PFNWGLGETSYNCVALUESOMLPROC __wglewGetSyncValuesOML;
WGLEW_EXPORT PFNWGLSWAPBUFFERSMSCOMLPROC __wglewSwapBuffersMscOML;
WGLEW_EXPORT PFNWGLSWAPLAYERBUFFERSMSCOMLPROC __wglewSwapLayerBuffersMscOML;
WGLEW_EXPORT PFNWGLWAITFORMSCOMLPROC __wglewWaitForMscOML;
WGLEW_EXPORT PFNWGLWAITFORSBCOMLPROC __wglewWaitForSbcOML;
WGLEW_EXPORT GLboolean __WGLEW_3DFX_multisample;
WGLEW_EXPORT GLboolean __WGLEW_3DL_stereo_control;
WGLEW_EXPORT GLboolean __WGLEW_ARB_buffer_region;
WGLEW_EXPORT GLboolean __WGLEW_ARB_extensions_string;
WGLEW_EXPORT GLboolean __WGLEW_ARB_make_current_read;
WGLEW_EXPORT GLboolean __WGLEW_ARB_multisample;
WGLEW_EXPORT GLboolean __WGLEW_ARB_pbuffer;
WGLEW_EXPORT GLboolean __WGLEW_ARB_pixel_format;
WGLEW_EXPORT GLboolean __WGLEW_ARB_pixel_format_float;
WGLEW_EXPORT GLboolean __WGLEW_ARB_render_texture;
WGLEW_EXPORT GLboolean __WGLEW_ATI_pixel_format_float;
WGLEW_EXPORT GLboolean __WGLEW_ATI_render_texture_rectangle;
WGLEW_EXPORT GLboolean __WGLEW_EXT_depth_float;
WGLEW_EXPORT GLboolean __WGLEW_EXT_display_color_table;
WGLEW_EXPORT GLboolean __WGLEW_EXT_extensions_string;
WGLEW_EXPORT GLboolean __WGLEW_EXT_framebuffer_sRGB;
WGLEW_EXPORT GLboolean __WGLEW_EXT_make_current_read;
WGLEW_EXPORT GLboolean __WGLEW_EXT_multisample;
WGLEW_EXPORT GLboolean __WGLEW_EXT_pbuffer;
WGLEW_EXPORT GLboolean __WGLEW_EXT_pixel_format;
WGLEW_EXPORT GLboolean __WGLEW_EXT_pixel_format_packed_float;
WGLEW_EXPORT GLboolean __WGLEW_EXT_swap_control;
WGLEW_EXPORT GLboolean __WGLEW_I3D_digital_video_control;
WGLEW_EXPORT GLboolean __WGLEW_I3D_gamma;
WGLEW_EXPORT GLboolean __WGLEW_I3D_genlock;
WGLEW_EXPORT GLboolean __WGLEW_I3D_image_buffer;
WGLEW_EXPORT GLboolean __WGLEW_I3D_swap_frame_lock;
WGLEW_EXPORT GLboolean __WGLEW_I3D_swap_frame_usage;
WGLEW_EXPORT GLboolean __WGLEW_NV_float_buffer;
WGLEW_EXPORT GLboolean __WGLEW_NV_render_depth_texture;
WGLEW_EXPORT GLboolean __WGLEW_NV_render_texture_rectangle;
WGLEW_EXPORT GLboolean __WGLEW_NV_vertex_array_range;
WGLEW_EXPORT GLboolean __WGLEW_OML_sync_control;
#ifdef GLEW_MX
}; /* WGLEWContextStruct */
#endif /* GLEW_MX */
/* ------------------------------------------------------------------------- */
#ifdef GLEW_MX
typedef struct WGLEWContextStruct WGLEWContext;
GLEWAPI GLenum wglewContextInit(WGLEWContext *ctx);
GLEWAPI GLboolean wglewContextIsSupported(WGLEWContext *ctx, const char *name);
#define wglewInit() wglewContextInit(wglewGetContext())
#define wglewIsSupported(x) wglewContextIsSupported(wglewGetContext(), x)
#define WGLEW_GET_VAR(x) wglewGetContext()->x
#define WGLEW_GET_FUN(x) wglewGetContext()->x
#else /* GLEW_MX */
#define WGLEW_GET_VAR(x) x
#define WGLEW_GET_FUN(x) x
GLEWAPI GLboolean wglewIsSupported(const char *name);
#endif /* GLEW_MX */
GLEWAPI GLboolean wglewGetExtension(const char *name);
#ifdef __cplusplus
}
#endif
#undef GLEWAPI
#endif /* __wglew_h__ */

696
Common/GL/wglext.h Normal file
View File

@@ -0,0 +1,696 @@
#ifndef __wglext_h_
#define __wglext_h_
#ifdef __cplusplus
extern "C" {
#endif
/*
** License Applicability. Except to the extent portions of this file are
** made subject to an alternative license as permitted in the SGI Free
** Software License B, Version 1.1 (the "License"), the contents of this
** file are subject only to the provisions of the License. You may not use
** this file except in compliance with the License. You may obtain a copy
** of the License at Silicon Graphics, Inc., attn: Legal Services, 1600
** Amphitheatre Parkway, Mountain View, CA 94043-1351, or at:
**
** http://oss.sgi.com/projects/FreeB
**
** Note that, as provided in the License, the Software is distributed on an
** "AS IS" basis, with ALL EXPRESS AND IMPLIED WARRANTIES AND CONDITIONS
** DISCLAIMED, INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTIES AND
** CONDITIONS OF MERCHANTABILITY, SATISFACTORY QUALITY, FITNESS FOR A
** PARTICULAR PURPOSE, AND NON-INFRINGEMENT.
**
** Original Code. The Original Code is: OpenGL Sample Implementation,
** Version 1.2.1, released January 26, 2000, developed by Silicon Graphics,
** Inc. The Original Code is Copyright (c) 1991-2004 Silicon Graphics, Inc.
** Copyright in any portions created by third parties is as indicated
** elsewhere herein. All Rights Reserved.
**
** Additional Notice Provisions: This software was created using the
** OpenGL(R) version 1.2.1 Sample Implementation published by SGI, but has
** not been independently verified as being compliant with the OpenGL(R)
** version 1.2.1 Specification.
*/
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64) && !defined(APIENTRY) && !defined(__CYGWIN__) && !defined(__SCITECH_SNAP__)
#define WIN32_LEAN_AND_MEAN 1
#include <windows.h>
#endif
#ifndef APIENTRY
#define APIENTRY
#endif
#ifndef APIENTRYP
#define APIENTRYP APIENTRY *
#endif
#ifndef GLAPI
#define GLAPI extern
#endif
/*************************************************************/
/* Header file version number */
/* wglext.h last updated 2005/01/07 */
/* Current version at http://oss.sgi.com/projects/ogl-sample/registry/ */
#define WGL_WGLEXT_VERSION 6
#ifndef WGL_ARB_buffer_region
#define WGL_FRONT_COLOR_BUFFER_BIT_ARB 0x00000001
#define WGL_BACK_COLOR_BUFFER_BIT_ARB 0x00000002
#define WGL_DEPTH_BUFFER_BIT_ARB 0x00000004
#define WGL_STENCIL_BUFFER_BIT_ARB 0x00000008
#endif
#ifndef WGL_ARB_multisample
#define WGL_SAMPLE_BUFFERS_ARB 0x2041
#define WGL_SAMPLES_ARB 0x2042
#endif
#ifndef WGL_ARB_extensions_string
#endif
#ifndef WGL_ARB_pixel_format
#define WGL_NUMBER_PIXEL_FORMATS_ARB 0x2000
#define WGL_DRAW_TO_WINDOW_ARB 0x2001
#define WGL_DRAW_TO_BITMAP_ARB 0x2002
#define WGL_ACCELERATION_ARB 0x2003
#define WGL_NEED_PALETTE_ARB 0x2004
#define WGL_NEED_SYSTEM_PALETTE_ARB 0x2005
#define WGL_SWAP_LAYER_BUFFERS_ARB 0x2006
#define WGL_SWAP_METHOD_ARB 0x2007
#define WGL_NUMBER_OVERLAYS_ARB 0x2008
#define WGL_NUMBER_UNDERLAYS_ARB 0x2009
#define WGL_TRANSPARENT_ARB 0x200A
#define WGL_TRANSPARENT_RED_VALUE_ARB 0x2037
#define WGL_TRANSPARENT_GREEN_VALUE_ARB 0x2038
#define WGL_TRANSPARENT_BLUE_VALUE_ARB 0x2039
#define WGL_TRANSPARENT_ALPHA_VALUE_ARB 0x203A
#define WGL_TRANSPARENT_INDEX_VALUE_ARB 0x203B
#define WGL_SHARE_DEPTH_ARB 0x200C
#define WGL_SHARE_STENCIL_ARB 0x200D
#define WGL_SHARE_ACCUM_ARB 0x200E
#define WGL_SUPPORT_GDI_ARB 0x200F
#define WGL_SUPPORT_OPENGL_ARB 0x2010
#define WGL_DOUBLE_BUFFER_ARB 0x2011
#define WGL_STEREO_ARB 0x2012
#define WGL_PIXEL_TYPE_ARB 0x2013
#define WGL_COLOR_BITS_ARB 0x2014
#define WGL_RED_BITS_ARB 0x2015
#define WGL_RED_SHIFT_ARB 0x2016
#define WGL_GREEN_BITS_ARB 0x2017
#define WGL_GREEN_SHIFT_ARB 0x2018
#define WGL_BLUE_BITS_ARB 0x2019
#define WGL_BLUE_SHIFT_ARB 0x201A
#define WGL_ALPHA_BITS_ARB 0x201B
#define WGL_ALPHA_SHIFT_ARB 0x201C
#define WGL_ACCUM_BITS_ARB 0x201D
#define WGL_ACCUM_RED_BITS_ARB 0x201E
#define WGL_ACCUM_GREEN_BITS_ARB 0x201F
#define WGL_ACCUM_BLUE_BITS_ARB 0x2020
#define WGL_ACCUM_ALPHA_BITS_ARB 0x2021
#define WGL_DEPTH_BITS_ARB 0x2022
#define WGL_STENCIL_BITS_ARB 0x2023
#define WGL_AUX_BUFFERS_ARB 0x2024
#define WGL_NO_ACCELERATION_ARB 0x2025
#define WGL_GENERIC_ACCELERATION_ARB 0x2026
#define WGL_FULL_ACCELERATION_ARB 0x2027
#define WGL_SWAP_EXCHANGE_ARB 0x2028
#define WGL_SWAP_COPY_ARB 0x2029
#define WGL_SWAP_UNDEFINED_ARB 0x202A
#define WGL_TYPE_RGBA_ARB 0x202B
#define WGL_TYPE_COLORINDEX_ARB 0x202C
#endif
#ifndef WGL_ARB_make_current_read
#define ERROR_INVALID_PIXEL_TYPE_ARB 0x2043
#define ERROR_INCOMPATIBLE_DEVICE_CONTEXTS_ARB 0x2054
#endif
#ifndef WGL_ARB_pbuffer
#define WGL_DRAW_TO_PBUFFER_ARB 0x202D
#define WGL_MAX_PBUFFER_PIXELS_ARB 0x202E
#define WGL_MAX_PBUFFER_WIDTH_ARB 0x202F
#define WGL_MAX_PBUFFER_HEIGHT_ARB 0x2030
#define WGL_PBUFFER_LARGEST_ARB 0x2033
#define WGL_PBUFFER_WIDTH_ARB 0x2034
#define WGL_PBUFFER_HEIGHT_ARB 0x2035
#define WGL_PBUFFER_LOST_ARB 0x2036
#endif
#ifndef WGL_ARB_render_texture
#define WGL_BIND_TO_TEXTURE_RGB_ARB 0x2070
#define WGL_BIND_TO_TEXTURE_RGBA_ARB 0x2071
#define WGL_TEXTURE_FORMAT_ARB 0x2072
#define WGL_TEXTURE_TARGET_ARB 0x2073
#define WGL_MIPMAP_TEXTURE_ARB 0x2074
#define WGL_TEXTURE_RGB_ARB 0x2075
#define WGL_TEXTURE_RGBA_ARB 0x2076
#define WGL_NO_TEXTURE_ARB 0x2077
#define WGL_TEXTURE_CUBE_MAP_ARB 0x2078
#define WGL_TEXTURE_1D_ARB 0x2079
#define WGL_TEXTURE_2D_ARB 0x207A
#define WGL_MIPMAP_LEVEL_ARB 0x207B
#define WGL_CUBE_MAP_FACE_ARB 0x207C
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_X_ARB 0x207D
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_X_ARB 0x207E
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_Y_ARB 0x207F
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_Y_ARB 0x2080
#define WGL_TEXTURE_CUBE_MAP_POSITIVE_Z_ARB 0x2081
#define WGL_TEXTURE_CUBE_MAP_NEGATIVE_Z_ARB 0x2082
#define WGL_FRONT_LEFT_ARB 0x2083
#define WGL_FRONT_RIGHT_ARB 0x2084
#define WGL_BACK_LEFT_ARB 0x2085
#define WGL_BACK_RIGHT_ARB 0x2086
#define WGL_AUX0_ARB 0x2087
#define WGL_AUX1_ARB 0x2088
#define WGL_AUX2_ARB 0x2089
#define WGL_AUX3_ARB 0x208A
#define WGL_AUX4_ARB 0x208B
#define WGL_AUX5_ARB 0x208C
#define WGL_AUX6_ARB 0x208D
#define WGL_AUX7_ARB 0x208E
#define WGL_AUX8_ARB 0x208F
#define WGL_AUX9_ARB 0x2090
#endif
#ifndef WGL_ARB_pixel_format_float
#define WGL_TYPE_RGBA_FLOAT_ARB 0x21A0
#endif
#ifndef WGL_EXT_make_current_read
#define ERROR_INVALID_PIXEL_TYPE_EXT 0x2043
#endif
#ifndef WGL_EXT_pixel_format
#define WGL_NUMBER_PIXEL_FORMATS_EXT 0x2000
#define WGL_DRAW_TO_WINDOW_EXT 0x2001
#define WGL_DRAW_TO_BITMAP_EXT 0x2002
#define WGL_ACCELERATION_EXT 0x2003
#define WGL_NEED_PALETTE_EXT 0x2004
#define WGL_NEED_SYSTEM_PALETTE_EXT 0x2005
#define WGL_SWAP_LAYER_BUFFERS_EXT 0x2006
#define WGL_SWAP_METHOD_EXT 0x2007
#define WGL_NUMBER_OVERLAYS_EXT 0x2008
#define WGL_NUMBER_UNDERLAYS_EXT 0x2009
#define WGL_TRANSPARENT_EXT 0x200A
#define WGL_TRANSPARENT_VALUE_EXT 0x200B
#define WGL_SHARE_DEPTH_EXT 0x200C
#define WGL_SHARE_STENCIL_EXT 0x200D
#define WGL_SHARE_ACCUM_EXT 0x200E
#define WGL_SUPPORT_GDI_EXT 0x200F
#define WGL_SUPPORT_OPENGL_EXT 0x2010
#define WGL_DOUBLE_BUFFER_EXT 0x2011
#define WGL_STEREO_EXT 0x2012
#define WGL_PIXEL_TYPE_EXT 0x2013
#define WGL_COLOR_BITS_EXT 0x2014
#define WGL_RED_BITS_EXT 0x2015
#define WGL_RED_SHIFT_EXT 0x2016
#define WGL_GREEN_BITS_EXT 0x2017
#define WGL_GREEN_SHIFT_EXT 0x2018
#define WGL_BLUE_BITS_EXT 0x2019
#define WGL_BLUE_SHIFT_EXT 0x201A
#define WGL_ALPHA_BITS_EXT 0x201B
#define WGL_ALPHA_SHIFT_EXT 0x201C
#define WGL_ACCUM_BITS_EXT 0x201D
#define WGL_ACCUM_RED_BITS_EXT 0x201E
#define WGL_ACCUM_GREEN_BITS_EXT 0x201F
#define WGL_ACCUM_BLUE_BITS_EXT 0x2020
#define WGL_ACCUM_ALPHA_BITS_EXT 0x2021
#define WGL_DEPTH_BITS_EXT 0x2022
#define WGL_STENCIL_BITS_EXT 0x2023
#define WGL_AUX_BUFFERS_EXT 0x2024
#define WGL_NO_ACCELERATION_EXT 0x2025
#define WGL_GENERIC_ACCELERATION_EXT 0x2026
#define WGL_FULL_ACCELERATION_EXT 0x2027
#define WGL_SWAP_EXCHANGE_EXT 0x2028
#define WGL_SWAP_COPY_EXT 0x2029
#define WGL_SWAP_UNDEFINED_EXT 0x202A
#define WGL_TYPE_RGBA_EXT 0x202B
#define WGL_TYPE_COLORINDEX_EXT 0x202C
#endif
#ifndef WGL_EXT_pbuffer
#define WGL_DRAW_TO_PBUFFER_EXT 0x202D
#define WGL_MAX_PBUFFER_PIXELS_EXT 0x202E
#define WGL_MAX_PBUFFER_WIDTH_EXT 0x202F
#define WGL_MAX_PBUFFER_HEIGHT_EXT 0x2030
#define WGL_OPTIMAL_PBUFFER_WIDTH_EXT 0x2031
#define WGL_OPTIMAL_PBUFFER_HEIGHT_EXT 0x2032
#define WGL_PBUFFER_LARGEST_EXT 0x2033
#define WGL_PBUFFER_WIDTH_EXT 0x2034
#define WGL_PBUFFER_HEIGHT_EXT 0x2035
#endif
#ifndef WGL_EXT_depth_float
#define WGL_DEPTH_FLOAT_EXT 0x2040
#endif
#ifndef WGL_3DFX_multisample
#define WGL_SAMPLE_BUFFERS_3DFX 0x2060
#define WGL_SAMPLES_3DFX 0x2061
#endif
#ifndef WGL_EXT_multisample
#define WGL_SAMPLE_BUFFERS_EXT 0x2041
#define WGL_SAMPLES_EXT 0x2042
#endif
#ifndef WGL_I3D_digital_video_control
#define WGL_DIGITAL_VIDEO_CURSOR_ALPHA_FRAMEBUFFER_I3D 0x2050
#define WGL_DIGITAL_VIDEO_CURSOR_ALPHA_VALUE_I3D 0x2051
#define WGL_DIGITAL_VIDEO_CURSOR_INCLUDED_I3D 0x2052
#define WGL_DIGITAL_VIDEO_GAMMA_CORRECTED_I3D 0x2053
#endif
#ifndef WGL_I3D_gamma
#define WGL_GAMMA_TABLE_SIZE_I3D 0x204E
#define WGL_GAMMA_EXCLUDE_DESKTOP_I3D 0x204F
#endif
#ifndef WGL_I3D_genlock
#define WGL_GENLOCK_SOURCE_MULTIVIEW_I3D 0x2044
#define WGL_GENLOCK_SOURCE_EXTENAL_SYNC_I3D 0x2045
#define WGL_GENLOCK_SOURCE_EXTENAL_FIELD_I3D 0x2046
#define WGL_GENLOCK_SOURCE_EXTENAL_TTL_I3D 0x2047
#define WGL_GENLOCK_SOURCE_DIGITAL_SYNC_I3D 0x2048
#define WGL_GENLOCK_SOURCE_DIGITAL_FIELD_I3D 0x2049
#define WGL_GENLOCK_SOURCE_EDGE_FALLING_I3D 0x204A
#define WGL_GENLOCK_SOURCE_EDGE_RISING_I3D 0x204B
#define WGL_GENLOCK_SOURCE_EDGE_BOTH_I3D 0x204C
#endif
#ifndef WGL_I3D_image_buffer
#define WGL_IMAGE_BUFFER_MIN_ACCESS_I3D 0x00000001
#define WGL_IMAGE_BUFFER_LOCK_I3D 0x00000002
#endif
#ifndef WGL_I3D_swap_frame_lock
#endif
#ifndef WGL_NV_render_depth_texture
#define WGL_BIND_TO_TEXTURE_DEPTH_NV 0x20A3
#define WGL_BIND_TO_TEXTURE_RECTANGLE_DEPTH_NV 0x20A4
#define WGL_DEPTH_TEXTURE_FORMAT_NV 0x20A5
#define WGL_TEXTURE_DEPTH_COMPONENT_NV 0x20A6
#define WGL_DEPTH_COMPONENT_NV 0x20A7
#endif
#ifndef WGL_NV_render_texture_rectangle
#define WGL_BIND_TO_TEXTURE_RECTANGLE_RGB_NV 0x20A0
#define WGL_BIND_TO_TEXTURE_RECTANGLE_RGBA_NV 0x20A1
#define WGL_TEXTURE_RECTANGLE_NV 0x20A2
#endif
#ifndef WGL_ATI_pixel_format_float
#define WGL_TYPE_RGBA_FLOAT_ATI 0x21A0
#define WGL_RGBA_FLOAT_MODE_ATI 0x8820
#define WGL_COLOR_CLEAR_UNCLAMPED_VALUE_ATI 0x8835
#endif
#ifndef WGL_NV_float_buffer
#define WGL_FLOAT_COMPONENTS_NV 0x20B0
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_R_NV 0x20B1
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RG_NV 0x20B2
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RGB_NV 0x20B3
#define WGL_BIND_TO_TEXTURE_RECTANGLE_FLOAT_RGBA_NV 0x20B4
#define WGL_TEXTURE_FLOAT_R_NV 0x20B5
#define WGL_TEXTURE_FLOAT_RG_NV 0x20B6
#define WGL_TEXTURE_FLOAT_RGB_NV 0x20B7
#define WGL_TEXTURE_FLOAT_RGBA_NV 0x20B8
#endif
#ifndef WGL_NV_swap_group
#endif
#ifndef WGL_NV_gpu_affinity
#define WGL_ERROR_INCOMPATIBLE_AFFINITY_MASKS_NV 0x20D0
#define WGL_ERROR_MISSING_AFFINITY_MASK_NV 0x20D1
#endif
/*************************************************************/
#ifndef WGL_ARB_pbuffer
DECLARE_HANDLE(HPBUFFERARB);
#endif
#ifndef WGL_EXT_pbuffer
DECLARE_HANDLE(HPBUFFEREXT);
#endif
#ifndef WGL_NV_gpu_affinity
DECLARE_HANDLE(HGPUNV);
typedef struct _GPU_DEVICE
{
DWORD cb;
CHAR DeviceName[32];
CHAR DeviceString[128];
DWORD Flags;
RECT rcVirtualScreen;
} GPU_DEVICE, *PGPU_DEVICE;
#endif
#ifndef WGL_ARB_buffer_region
#define WGL_ARB_buffer_region 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern HANDLE WINAPI wglCreateBufferRegionARB(HDC, int, UINT);
extern VOID WINAPI wglDeleteBufferRegionARB(HANDLE);
extern BOOL WINAPI wglSaveBufferRegionARB(HANDLE, int, int, int, int);
extern BOOL WINAPI wglRestoreBufferRegionARB(HANDLE, int, int, int, int, int, int);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef HANDLE(WINAPI *PFNWGLCREATEBUFFERREGIONARBPROC)(HDC hDC, int iLayerPlane, UINT uType);
typedef VOID (WINAPI *PFNWGLDELETEBUFFERREGIONARBPROC)(HANDLE hRegion);
typedef BOOL (WINAPI *PFNWGLSAVEBUFFERREGIONARBPROC)(HANDLE hRegion, int x, int y, int width, int height);
typedef BOOL (WINAPI *PFNWGLRESTOREBUFFERREGIONARBPROC)(HANDLE hRegion, int x, int y, int width, int height, int xSrc, int ySrc);
#endif
#ifndef WGL_ARB_multisample
#define WGL_ARB_multisample 1
#endif
#ifndef WGL_ARB_extensions_string
#define WGL_ARB_extensions_string 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern const char *WINAPI wglGetExtensionsStringARB(HDC);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef const char *(WINAPI *PFNWGLGETEXTENSIONSSTRINGARBPROC)(HDC hdc);
#endif
#ifndef WGL_ARB_pixel_format
#define WGL_ARB_pixel_format 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetPixelFormatAttribivARB(HDC, int, int, UINT, const int *, int *);
extern BOOL WINAPI wglGetPixelFormatAttribfvARB(HDC, int, int, UINT, const int *, FLOAT *);
extern BOOL WINAPI wglChoosePixelFormatARB(HDC, const int *, const FLOAT *, UINT, int *, UINT *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBIVARBPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, const int *piAttributes, int *piValues);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBFVARBPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, const int *piAttributes, FLOAT *pfValues);
typedef BOOL (WINAPI *PFNWGLCHOOSEPIXELFORMATARBPROC)(HDC hdc, const int *piAttribIList, const FLOAT *pfAttribFList, UINT nMaxFormats, int *piFormats, UINT *nNumFormats);
#endif
#ifndef WGL_ARB_make_current_read
#define WGL_ARB_make_current_read 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglMakeContextCurrentARB(HDC, HDC, HGLRC);
extern HDC WINAPI wglGetCurrentReadDCARB(void);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLMAKECONTEXTCURRENTARBPROC)(HDC hDrawDC, HDC hReadDC, HGLRC hglrc);
typedef HDC(WINAPI *PFNWGLGETCURRENTREADDCARBPROC)(void);
#endif
#ifndef WGL_ARB_pbuffer
#define WGL_ARB_pbuffer 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern HPBUFFERARB WINAPI wglCreatePbufferARB(HDC, int, int, int, const int *);
extern HDC WINAPI wglGetPbufferDCARB(HPBUFFERARB);
extern int WINAPI wglReleasePbufferDCARB(HPBUFFERARB, HDC);
extern BOOL WINAPI wglDestroyPbufferARB(HPBUFFERARB);
extern BOOL WINAPI wglQueryPbufferARB(HPBUFFERARB, int, int *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef HPBUFFERARB(WINAPI *PFNWGLCREATEPBUFFERARBPROC)(HDC hDC, int iPixelFormat, int iWidth, int iHeight, const int *piAttribList);
typedef HDC(WINAPI *PFNWGLGETPBUFFERDCARBPROC)(HPBUFFERARB hPbuffer);
typedef int (WINAPI *PFNWGLRELEASEPBUFFERDCARBPROC)(HPBUFFERARB hPbuffer, HDC hDC);
typedef BOOL (WINAPI *PFNWGLDESTROYPBUFFERARBPROC)(HPBUFFERARB hPbuffer);
typedef BOOL (WINAPI *PFNWGLQUERYPBUFFERARBPROC)(HPBUFFERARB hPbuffer, int iAttribute, int *piValue);
#endif
#ifndef WGL_ARB_render_texture
#define WGL_ARB_render_texture 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglBindTexImageARB(HPBUFFERARB, int);
extern BOOL WINAPI wglReleaseTexImageARB(HPBUFFERARB, int);
extern BOOL WINAPI wglSetPbufferAttribARB(HPBUFFERARB, const int *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLBINDTEXIMAGEARBPROC)(HPBUFFERARB hPbuffer, int iBuffer);
typedef BOOL (WINAPI *PFNWGLRELEASETEXIMAGEARBPROC)(HPBUFFERARB hPbuffer, int iBuffer);
typedef BOOL (WINAPI *PFNWGLSETPBUFFERATTRIBARBPROC)(HPBUFFERARB hPbuffer, const int *piAttribList);
#endif
#ifndef WGL_ARB_pixel_format_float
#define WGL_ARB_pixel_format_float 1
#endif
#ifndef WGL_EXT_display_color_table
#define WGL_EXT_display_color_table 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern GLboolean WINAPI wglCreateDisplayColorTableEXT(GLushort);
extern GLboolean WINAPI wglLoadDisplayColorTableEXT(const GLushort *, GLuint);
extern GLboolean WINAPI wglBindDisplayColorTableEXT(GLushort);
extern VOID WINAPI wglDestroyDisplayColorTableEXT(GLushort);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef GLboolean(WINAPI *PFNWGLCREATEDISPLAYCOLORTABLEEXTPROC)(GLushort id);
typedef GLboolean(WINAPI *PFNWGLLOADDISPLAYCOLORTABLEEXTPROC)(const GLushort *table, GLuint length);
typedef GLboolean(WINAPI *PFNWGLBINDDISPLAYCOLORTABLEEXTPROC)(GLushort id);
typedef VOID (WINAPI *PFNWGLDESTROYDISPLAYCOLORTABLEEXTPROC)(GLushort id);
#endif
#ifndef WGL_EXT_extensions_string
#define WGL_EXT_extensions_string 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern const char *WINAPI wglGetExtensionsStringEXT(void);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef const char *(WINAPI *PFNWGLGETEXTENSIONSSTRINGEXTPROC)(void);
#endif
#ifndef WGL_EXT_make_current_read
#define WGL_EXT_make_current_read 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglMakeContextCurrentEXT(HDC, HDC, HGLRC);
extern HDC WINAPI wglGetCurrentReadDCEXT(void);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLMAKECONTEXTCURRENTEXTPROC)(HDC hDrawDC, HDC hReadDC, HGLRC hglrc);
typedef HDC(WINAPI *PFNWGLGETCURRENTREADDCEXTPROC)(void);
#endif
#ifndef WGL_EXT_pbuffer
#define WGL_EXT_pbuffer 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern HPBUFFEREXT WINAPI wglCreatePbufferEXT(HDC, int, int, int, const int *);
extern HDC WINAPI wglGetPbufferDCEXT(HPBUFFEREXT);
extern int WINAPI wglReleasePbufferDCEXT(HPBUFFEREXT, HDC);
extern BOOL WINAPI wglDestroyPbufferEXT(HPBUFFEREXT);
extern BOOL WINAPI wglQueryPbufferEXT(HPBUFFEREXT, int, int *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef HPBUFFEREXT(WINAPI *PFNWGLCREATEPBUFFEREXTPROC)(HDC hDC, int iPixelFormat, int iWidth, int iHeight, const int *piAttribList);
typedef HDC(WINAPI *PFNWGLGETPBUFFERDCEXTPROC)(HPBUFFEREXT hPbuffer);
typedef int (WINAPI *PFNWGLRELEASEPBUFFERDCEXTPROC)(HPBUFFEREXT hPbuffer, HDC hDC);
typedef BOOL (WINAPI *PFNWGLDESTROYPBUFFEREXTPROC)(HPBUFFEREXT hPbuffer);
typedef BOOL (WINAPI *PFNWGLQUERYPBUFFEREXTPROC)(HPBUFFEREXT hPbuffer, int iAttribute, int *piValue);
#endif
#ifndef WGL_EXT_pixel_format
#define WGL_EXT_pixel_format 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetPixelFormatAttribivEXT(HDC, int, int, UINT, int *, int *);
extern BOOL WINAPI wglGetPixelFormatAttribfvEXT(HDC, int, int, UINT, int *, FLOAT *);
extern BOOL WINAPI wglChoosePixelFormatEXT(HDC, const int *, const FLOAT *, UINT, int *, UINT *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBIVEXTPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, int *piAttributes, int *piValues);
typedef BOOL (WINAPI *PFNWGLGETPIXELFORMATATTRIBFVEXTPROC)(HDC hdc, int iPixelFormat, int iLayerPlane, UINT nAttributes, int *piAttributes, FLOAT *pfValues);
typedef BOOL (WINAPI *PFNWGLCHOOSEPIXELFORMATEXTPROC)(HDC hdc, const int *piAttribIList, const FLOAT *pfAttribFList, UINT nMaxFormats, int *piFormats, UINT *nNumFormats);
#endif
#ifndef WGL_EXT_swap_control
#define WGL_EXT_swap_control 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglSwapIntervalEXT(int);
extern int WINAPI wglGetSwapIntervalEXT(void);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLSWAPINTERVALEXTPROC)(int interval);
typedef int (WINAPI *PFNWGLGETSWAPINTERVALEXTPROC)(void);
#endif
#ifndef WGL_EXT_depth_float
#define WGL_EXT_depth_float 1
#endif
#ifndef WGL_NV_vertex_array_range
#define WGL_NV_vertex_array_range 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern void *WINAPI wglAllocateMemoryNV(GLsizei, GLfloat, GLfloat, GLfloat);
extern void WINAPI wglFreeMemoryNV(void *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef void *(WINAPI *PFNWGLALLOCATEMEMORYNVPROC)(GLsizei size, GLfloat readfreq, GLfloat writefreq, GLfloat priority);
typedef void (WINAPI *PFNWGLFREEMEMORYNVPROC)(void *pointer);
#endif
#ifndef WGL_3DFX_multisample
#define WGL_3DFX_multisample 1
#endif
#ifndef WGL_EXT_multisample
#define WGL_EXT_multisample 1
#endif
#ifndef WGL_OML_sync_control
#define WGL_OML_sync_control 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetSyncValuesOML(HDC, INT64 *, INT64 *, INT64 *);
extern BOOL WINAPI wglGetMscRateOML(HDC, INT32 *, INT32 *);
extern INT64 WINAPI wglSwapBuffersMscOML(HDC, INT64, INT64, INT64);
extern INT64 WINAPI wglSwapLayerBuffersMscOML(HDC, int, INT64, INT64, INT64);
extern BOOL WINAPI wglWaitForMscOML(HDC, INT64, INT64, INT64, INT64 *, INT64 *, INT64 *);
extern BOOL WINAPI wglWaitForSbcOML(HDC, INT64, INT64 *, INT64 *, INT64 *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETSYNCVALUESOMLPROC)(HDC hdc, INT64 *ust, INT64 *msc, INT64 *sbc);
typedef BOOL (WINAPI *PFNWGLGETMSCRATEOMLPROC)(HDC hdc, INT32 *numerator, INT32 *denominator);
typedef INT64(WINAPI *PFNWGLSWAPBUFFERSMSCOMLPROC)(HDC hdc, INT64 target_msc, INT64 divisor, INT64 remainder);
typedef INT64(WINAPI *PFNWGLSWAPLAYERBUFFERSMSCOMLPROC)(HDC hdc, int fuPlanes, INT64 target_msc, INT64 divisor, INT64 remainder);
typedef BOOL (WINAPI *PFNWGLWAITFORMSCOMLPROC)(HDC hdc, INT64 target_msc, INT64 divisor, INT64 remainder, INT64 *ust, INT64 *msc, INT64 *sbc);
typedef BOOL (WINAPI *PFNWGLWAITFORSBCOMLPROC)(HDC hdc, INT64 target_sbc, INT64 *ust, INT64 *msc, INT64 *sbc);
#endif
#ifndef WGL_I3D_digital_video_control
#define WGL_I3D_digital_video_control 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetDigitalVideoParametersI3D(HDC, int, int *);
extern BOOL WINAPI wglSetDigitalVideoParametersI3D(HDC, int, const int *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETDIGITALVIDEOPARAMETERSI3DPROC)(HDC hDC, int iAttribute, int *piValue);
typedef BOOL (WINAPI *PFNWGLSETDIGITALVIDEOPARAMETERSI3DPROC)(HDC hDC, int iAttribute, const int *piValue);
#endif
#ifndef WGL_I3D_gamma
#define WGL_I3D_gamma 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetGammaTableParametersI3D(HDC, int, int *);
extern BOOL WINAPI wglSetGammaTableParametersI3D(HDC, int, const int *);
extern BOOL WINAPI wglGetGammaTableI3D(HDC, int, USHORT *, USHORT *, USHORT *);
extern BOOL WINAPI wglSetGammaTableI3D(HDC, int, const USHORT *, const USHORT *, const USHORT *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETGAMMATABLEPARAMETERSI3DPROC)(HDC hDC, int iAttribute, int *piValue);
typedef BOOL (WINAPI *PFNWGLSETGAMMATABLEPARAMETERSI3DPROC)(HDC hDC, int iAttribute, const int *piValue);
typedef BOOL (WINAPI *PFNWGLGETGAMMATABLEI3DPROC)(HDC hDC, int iEntries, USHORT *puRed, USHORT *puGreen, USHORT *puBlue);
typedef BOOL (WINAPI *PFNWGLSETGAMMATABLEI3DPROC)(HDC hDC, int iEntries, const USHORT *puRed, const USHORT *puGreen, const USHORT *puBlue);
#endif
#ifndef WGL_I3D_genlock
#define WGL_I3D_genlock 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglEnableGenlockI3D(HDC);
extern BOOL WINAPI wglDisableGenlockI3D(HDC);
extern BOOL WINAPI wglIsEnabledGenlockI3D(HDC, BOOL *);
extern BOOL WINAPI wglGenlockSourceI3D(HDC, UINT);
extern BOOL WINAPI wglGetGenlockSourceI3D(HDC, UINT *);
extern BOOL WINAPI wglGenlockSourceEdgeI3D(HDC, UINT);
extern BOOL WINAPI wglGetGenlockSourceEdgeI3D(HDC, UINT *);
extern BOOL WINAPI wglGenlockSampleRateI3D(HDC, UINT);
extern BOOL WINAPI wglGetGenlockSampleRateI3D(HDC, UINT *);
extern BOOL WINAPI wglGenlockSourceDelayI3D(HDC, UINT);
extern BOOL WINAPI wglGetGenlockSourceDelayI3D(HDC, UINT *);
extern BOOL WINAPI wglQueryGenlockMaxSourceDelayI3D(HDC, UINT *, UINT *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLENABLEGENLOCKI3DPROC)(HDC hDC);
typedef BOOL (WINAPI *PFNWGLDISABLEGENLOCKI3DPROC)(HDC hDC);
typedef BOOL (WINAPI *PFNWGLISENABLEDGENLOCKI3DPROC)(HDC hDC, BOOL *pFlag);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEI3DPROC)(HDC hDC, UINT uSource);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEI3DPROC)(HDC hDC, UINT *uSource);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEEDGEI3DPROC)(HDC hDC, UINT uEdge);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEEDGEI3DPROC)(HDC hDC, UINT *uEdge);
typedef BOOL (WINAPI *PFNWGLGENLOCKSAMPLERATEI3DPROC)(HDC hDC, UINT uRate);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSAMPLERATEI3DPROC)(HDC hDC, UINT *uRate);
typedef BOOL (WINAPI *PFNWGLGENLOCKSOURCEDELAYI3DPROC)(HDC hDC, UINT uDelay);
typedef BOOL (WINAPI *PFNWGLGETGENLOCKSOURCEDELAYI3DPROC)(HDC hDC, UINT *uDelay);
typedef BOOL (WINAPI *PFNWGLQUERYGENLOCKMAXSOURCEDELAYI3DPROC)(HDC hDC, UINT *uMaxLineDelay, UINT *uMaxPixelDelay);
#endif
#ifndef WGL_I3D_image_buffer
#define WGL_I3D_image_buffer 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern LPVOID WINAPI wglCreateImageBufferI3D(HDC, DWORD, UINT);
extern BOOL WINAPI wglDestroyImageBufferI3D(HDC, LPVOID);
extern BOOL WINAPI wglAssociateImageBufferEventsI3D(HDC, const HANDLE *, const LPVOID *, const DWORD *, UINT);
extern BOOL WINAPI wglReleaseImageBufferEventsI3D(HDC, const LPVOID *, UINT);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef LPVOID (WINAPI *PFNWGLCREATEIMAGEBUFFERI3DPROC)(HDC hDC, DWORD dwSize, UINT uFlags);
typedef BOOL (WINAPI *PFNWGLDESTROYIMAGEBUFFERI3DPROC)(HDC hDC, LPVOID pAddress);
typedef BOOL (WINAPI *PFNWGLASSOCIATEIMAGEBUFFEREVENTSI3DPROC)(HDC hDC, const HANDLE *pEvent, const LPVOID *pAddress, const DWORD *pSize, UINT count);
typedef BOOL (WINAPI *PFNWGLRELEASEIMAGEBUFFEREVENTSI3DPROC)(HDC hDC, const LPVOID *pAddress, UINT count);
#endif
#ifndef WGL_I3D_swap_frame_lock
#define WGL_I3D_swap_frame_lock 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglEnableFrameLockI3D(void);
extern BOOL WINAPI wglDisableFrameLockI3D(void);
extern BOOL WINAPI wglIsEnabledFrameLockI3D(BOOL *);
extern BOOL WINAPI wglQueryFrameLockMasterI3D(BOOL *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLENABLEFRAMELOCKI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLDISABLEFRAMELOCKI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLISENABLEDFRAMELOCKI3DPROC)(BOOL *pFlag);
typedef BOOL (WINAPI *PFNWGLQUERYFRAMELOCKMASTERI3DPROC)(BOOL *pFlag);
#endif
#ifndef WGL_I3D_swap_frame_usage
#define WGL_I3D_swap_frame_usage 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglGetFrameUsageI3D(float *);
extern BOOL WINAPI wglBeginFrameTrackingI3D(void);
extern BOOL WINAPI wglEndFrameTrackingI3D(void);
extern BOOL WINAPI wglQueryFrameTrackingI3D(DWORD *, DWORD *, float *);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLGETFRAMEUSAGEI3DPROC)(float *pUsage);
typedef BOOL (WINAPI *PFNWGLBEGINFRAMETRACKINGI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLENDFRAMETRACKINGI3DPROC)(void);
typedef BOOL (WINAPI *PFNWGLQUERYFRAMETRACKINGI3DPROC)(DWORD *pFrameCount, DWORD *pMissedFrames, float *pLastMissedUsage);
#endif
#ifndef WGL_ATI_pixel_format_float
#define WGL_ATI_pixel_format_float 1
#endif
#ifndef WGL_NV_render_depth_texture
#define WGL_NV_render_depth_texture 1
#endif
#ifndef WGL_NV_render_texture_rectangle
#define WGL_NV_render_texture_rectangle 1
#endif
#ifndef WGL_NV_float_buffer
#define WGL_NV_float_buffer 1
#endif
#ifndef WGL_NV_swap_group
#define WGL_NV_swap_group 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglJoinSwapGroupNV(HDC hDC, GLuint group);
extern BOOL WINAPI wglBindSwapBarrierNV(GLuint group, GLuint barrier);
extern BOOL WINAPI wglQuerySwapGroupNV(HDC hDC, GLuint *group, GLuint *barrier);
extern BOOL WINAPI wglQueryMaxSwapGroupsNV(HDC hDC, GLuint *maxGroups, GLuint *maxBarriers);
extern BOOL WINAPI wglQueryFrameCountNV(HDC hDC, GLuint *count);
extern BOOL WINAPI wglResetFrameCountNV(HDC hDC);
#endif /* WGL_WGLEXT_PROTOTYPES */
typedef BOOL (WINAPI *PFNWGLJOINSWAPGROUPNVPROC)(HDC hDC, GLuint group);
typedef BOOL (WINAPI *PFNWGLBINDSWAPBARRIERNVPROC)(GLuint group, GLuint barrier);
typedef BOOL (WINAPI *PFNWGLQUERYSWAPGROUPNVPROC)(HDC hDC, GLuint *group, GLuint *barrier);
typedef BOOL (WINAPI *PFNWGLQUERYMAXSWAPGROUPSNVPROC)(HDC hDC, GLuint *maxGroups, GLuint *maxBarriers);
typedef BOOL (WINAPI *PFNWGLQUERYFRAMECOUNTNVPROC)(HDC hDC, GLuint *count);
typedef BOOL (WINAPI *PFNWGLRESETFRAMECOUNTNVPROC)(HDC hDC);
#endif
#ifndef WGL_NV_gpu_affinity
#define WGL_NV_gpu_affinity 1
#ifdef WGL_WGLEXT_PROTOTYPES
extern BOOL WINAPI wglEnumGpusNV(UINT iIndex, HGPUNV *hGpu);
extern BOOL WINAPI wglEnumGpuDevicesNV(HGPUNV hGpu, UINT iIndex, PGPU_DEVICE pGpuDevice);
extern HDC WINAPI wglCreateAffinityDCNV(const HGPUNV *pGpuList);
extern BOOL WINAPI wglEnumGpusFromAffinityDCNV(HDC hAffinityDC, UINT iIndex, HGPUNV *hGpu);
extern BOOL WINAPI wglDeleteDCNV(HDC hAffinityDC);
#else
typedef BOOL (WINAPI *PFNWGLENUMGPUSNVPROC)(UINT iIndex, HGPUNV *hGpu);
typedef BOOL (WINAPI *PFNWGLENUMGPUDEVICESNVPROC)(HGPUNV hGpu, UINT iIndex, PGPU_DEVICE pGpuDevice);
typedef HDC(WINAPI *PFNWGLCREATEAFFINITYDCNVPROC)(const HGPUNV *pGpuList);
typedef BOOL (WINAPI *PFNWGLENUMGPUSFROMAFFINITYDCNVPROC)(HDC hAffinityDC, UINT iIndex, HGPUNV *hGpu);
typedef BOOL (WINAPI *PFNWGLDELETEDCNVPROC)(HDC hAffinityDC);
#endif
#endif
#ifdef __cplusplus
}
#endif
#endif

197
Common/UtilNPP/Exceptions.h Normal file
View File

@@ -0,0 +1,197 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_EXCEPTIONS_H
#define NV_UTIL_NPP_EXCEPTIONS_H
#include <string>
#include <sstream>
#include <iostream>
/// All npp related C++ classes are put into the npp namespace.
namespace npp
{
/// Exception base class.
/// This exception base class will be used for everything C++ throught
/// the NPP project.
/// The exception contains a string message, as well as data fields for a string
/// containing the name of the file as well as the line number where the exception was thrown.
/// The easiest way of throwing exceptions and providing filename and line number is
/// to use one of the ASSERT macros defined for that purpose.
class Exception
{
public:
/// Constructor.
/// \param rMessage A message with information as to why the exception was thrown.
/// \param rFileName The name of the file where the exception was thrown.
/// \param nLineNumber Line number in the file where the exception was thrown.
explicit
Exception(const std::string &rMessage = "", const std::string &rFileName = "", unsigned int nLineNumber = 0)
: sMessage_(rMessage), sFileName_(rFileName), nLineNumber_(nLineNumber)
{ };
Exception(const Exception &rException)
: sMessage_(rException.sMessage_), sFileName_(rException.sFileName_), nLineNumber_(rException.nLineNumber_)
{ };
virtual
~Exception()
{ };
/// Get the exception's message.
const
std::string &
message()
const
{
return sMessage_;
}
/// Get the exception's file info.
const
std::string &
fileName()
const
{
return sFileName_;
}
/// Get the exceptions's line info.
unsigned int
lineNumber()
const
{
return nLineNumber_;
}
/// Create a clone of this exception.
/// This creates a new Exception object on the heap. It is
/// the responsibility of the user of this function to free this memory
/// (delete x).
virtual
Exception *
clone()
const
{
return new Exception(*this);
}
/// Create a single string with all the exceptions information.
/// The virtual toString() method is used by the operator<<()
/// so that all exceptions derived from this base-class can print
/// their full information correctly even if a reference to their
/// exact type is not had at the time of printing (i.e. the basic
/// operator<<() is used).
virtual
std::string
toString()
const
{
std::ostringstream oOutputString;
oOutputString << fileName() << ":" << lineNumber() << ": " << message();
return oOutputString.str();
}
private:
std::string sMessage_; ///< Message regarding the cause of the exception.
std::string sFileName_; ///< Name of the file where the exception was thrown.
unsigned int nLineNumber_; ///< Line number in the file where the exception was thrown
};
/// Output stream inserter for Exception.
/// \param rOutputStream The stream the exception information is written to.
/// \param rException The exception that's being written.
/// \return Reference to the output stream being used.
std::ostream &
operator << (std::ostream &rOutputStream, const Exception &rException)
{
rOutputStream << rException.toString();
return rOutputStream;
}
/// Basic assert macro.
/// This macro should be used to enforce any kind of pre or post conditions.
/// Unlike the C-runtime assert macro, this macro does not abort execution, but throws
/// a C++ exception. The exception is automatically filled with information about the failing
/// condition, the filename and line number where the exception was thrown.
/// \note The macro is written in such a way that omitting a semicolon after its usage
/// causes a compiler error. The correct way to invoke this macro is:
/// NPP_ASSERT(n < MAX);
#define NPP_ASSERT(C) do {if (!(C)) throw npp::Exception(#C " assertion faild!", __FILE__, __LINE__);} while(false)
// ASSERT macro.
// Same functionality as the basic assert macro with the added ability to pass
// a message M. M should be a string literal.
// Note: Never use code inside ASSERT() that causes a side-effect ASSERT macros may get compiled
// out in release mode.
#define NPP_ASSERT_MSG(C, M) do {if (!(C)) throw npp::Exception(#C " assertion faild! Message: " M, __FILE__, __LINE__);} while(false)
#ifdef _DEBUG
/// Basic debug assert macro.
/// This macro is identical in every respect to NPP_ASSERT(C) but it does get compiled to a
/// no-op in release builds. It is therefor of utmost importance to not put statements into
/// this macro that cause side effects required for correct program execution.
#define NPP_DEBUG_ASSERT(C) do {if (!(C)) throw npp::Exception(#C " debug assertion faild!", __FILE__, __LINE__);} while(false)
#else
#define NPP_DEBUG_ASSERT(C)
#endif
/// ASSERT for null-pointer test.
/// It is safe to put code with side effects into this macro. Also: This macro never
/// gets compiled to a no-op because resource allocation may fail based on external causes not under
/// control of a software developer.
#define NPP_ASSERT_NOT_NULL(P) do {if ((P) == 0) throw npp::Exception(#P " not null assertion faild!", __FILE__, __LINE__);} while(false)
/// Macro for flagging methods as not implemented.
/// The macro throws an exception with a message that an implementation was missing
#define NPP_NOT_IMPLEMENTED() do {throw npp::Exception("Implementation missing!", __FILE__, __LINE__);} while(false)
/// Macro for checking error return code of CUDA (runtime) calls.
/// This macro never gets disabled.
#define NPP_CHECK_CUDA(S) do {cudaError_t eCUDAResult; \
eCUDAResult = S; \
if (eCUDAResult != cudaSuccess) std::cout << "NPP_CHECK_CUDA - eCUDAResult = " << eCUDAResult << std::endl; \
NPP_ASSERT(eCUDAResult == cudaSuccess);} while (false)
/// Macro for checking error return code for NPP calls.
#define NPP_CHECK_NPP(S) do {NppStatus eStatusNPP; \
eStatusNPP = S; \
if (eStatusNPP != NPP_SUCCESS) std::cout << "NPP_CHECK_NPP - eStatusNPP = " << _cudaGetErrorEnum(eStatusNPP) << "("<< eStatusNPP << ")" << std::endl; \
NPP_ASSERT(eStatusNPP == NPP_SUCCESS);} while (false)
/// Macro for checking error return codes from cuFFT calls.
#define NPP_CHECK_CUFFT(S) do {cufftResult eCUFFTResult; \
eCUFFTResult = S; \
if (eCUFFTResult != NPP_SUCCESS) std::cout << "NPP_CHECK_CUFFT - eCUFFTResult = " << eCUFFTResult << std::endl; \
NPP_ASSERT(eCUFFTResult == CUFFT_SUCCESS);} while (false)
} // npp namespace
#endif // NV_UTIL_NPP_EXCEPTIONS_H

155
Common/UtilNPP/Image.h Normal file
View File

@@ -0,0 +1,155 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGE_H
#define NV_UTIL_NPP_IMAGE_H
#include <cstddef>
namespace npp
{
class Image
{
public:
struct Size
{
unsigned int nWidth;
unsigned int nHeight;
Size() : nWidth(0), nHeight(0)
{ };
Size(unsigned int nWidthNew, unsigned nHeightNew) : nWidth(nWidthNew), nHeight(nHeightNew)
{ };
Size(const Size &rSize) : nWidth(rSize.nWidth), nHeight(rSize.nHeight)
{ };
Size &
operator= (const Size &rSize)
{
if (&rSize == this)
{
return *this;
}
nWidth = rSize.nWidth;
nHeight = rSize.nHeight;
return *this;
}
void
swap(Size &rSize)
{
unsigned int nTemp;
nTemp = nWidth;
nWidth = rSize.nWidth;
rSize.nWidth = nTemp;
nTemp = nHeight;
nHeight = rSize.nHeight;
rSize.nHeight = nTemp;
}
};
Image()
{ };
Image(unsigned int nWidth, unsigned int nHeight) : oSize_(nWidth, nHeight)
{ };
Image(const Image::Size &rSize) : oSize_(rSize)
{ };
Image(const Image &rImage) : oSize_(rImage.oSize_)
{ };
virtual
~Image()
{ };
Image &
operator= (const Image &rImage)
{
if (&rImage == this)
{
return *this;
}
oSize_ = rImage.oSize_;
return *this;
};
unsigned int
width()
const
{
return oSize_.nWidth;
}
unsigned int
height()
const
{
return oSize_.nHeight;
}
Size
size()
const
{
return oSize_;
}
void
swap(Image &rImage)
{
oSize_.swap(rImage.oSize_);
}
private:
Size oSize_;
};
bool
operator== (const Image::Size &rFirst, const Image::Size &rSecond)
{
return rFirst.nWidth == rSecond.nWidth && rFirst.nHeight == rSecond.nHeight;
}
bool
operator!= (const Image::Size &rFirst, const Image::Size &rSecond)
{
return rFirst.nWidth != rSecond.nWidth || rFirst.nHeight != rSecond.nHeight;
}
} // npp namespace
#endif // NV_UTIL_NPP_IMAGE_H

View File

@@ -0,0 +1,80 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGE_ALLOCATORS_CPU_H
#define NV_UTIL_NPP_IMAGE_ALLOCATORS_CPU_H
#include "Exceptions.h"
namespace npp
{
template <typename D, size_t N>
class ImageAllocatorCPU
{
public:
static
D *
Malloc2D(unsigned int nWidth, unsigned int nHeight, unsigned int *pPitch)
{
NPP_ASSERT(nWidth * nHeight > 0);
D *pResult = new D[nWidth * N * nHeight];
*pPitch = nWidth * sizeof(D) * N;
return pResult;
};
static
void
Free2D(D *pPixels)
{
delete[] pPixels;
};
static
void
Copy2D(D *pDst, size_t nDstPitch, const D *pSrc, size_t nSrcPitch, size_t nWidth, size_t nHeight)
{
const void *pSrcLine = pSrc;
void *pDstLine = pDst;
for (size_t iLine = 0; iLine < nHeight; ++iLine)
{
// copy one line worth of data
memcpy(pDst, pSrc, nWidth * N * sizeof(D));
// move data pointers to next line
pDst += nDstPitch;
pSrc += nSrcPitch;
}
};
};
} // npp namespace
#endif // NV_UTIL_NPP_IMAGE_ALLOCATORS_CPU_H

File diff suppressed because it is too large Load Diff

149
Common/UtilNPP/ImageIO.h Normal file
View File

@@ -0,0 +1,149 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGE_IO_H
#define NV_UTIL_NPP_IMAGE_IO_H
#include "ImagesCPU.h"
#include "ImagesNPP.h"
#include "FreeImage.h"
#include "Exceptions.h"
#include <string>
#include "string.h"
// Error handler for FreeImage library.
// In case this handler is invoked, it throws an NPP exception.
void
FreeImageErrorHandler(FREE_IMAGE_FORMAT oFif, const char *zMessage)
{
throw npp::Exception(zMessage);
}
namespace npp
{
// Load a gray-scale image from disk.
void
loadImage(const std::string &rFileName, ImageCPU_8u_C1 &rImage)
{
// set your own FreeImage error handler
FreeImage_SetOutputMessage(FreeImageErrorHandler);
FREE_IMAGE_FORMAT eFormat = FreeImage_GetFileType(rFileName.c_str());
// no signature? try to guess the file format from the file extension
if (eFormat == FIF_UNKNOWN)
{
eFormat = FreeImage_GetFIFFromFilename(rFileName.c_str());
}
NPP_ASSERT(eFormat != FIF_UNKNOWN);
// check that the plugin has reading capabilities ...
FIBITMAP *pBitmap;
if (FreeImage_FIFSupportsReading(eFormat))
{
pBitmap = FreeImage_Load(eFormat, rFileName.c_str());
}
NPP_ASSERT(pBitmap != 0);
// make sure this is an 8-bit single channel image
NPP_ASSERT(FreeImage_GetColorType(pBitmap) == FIC_MINISBLACK);
NPP_ASSERT(FreeImage_GetBPP(pBitmap) == 8);
// create an ImageCPU to receive the loaded image data
ImageCPU_8u_C1 oImage(FreeImage_GetWidth(pBitmap), FreeImage_GetHeight(pBitmap));
// Copy the FreeImage data into the new ImageCPU
unsigned int nSrcPitch = FreeImage_GetPitch(pBitmap);
const Npp8u *pSrcLine = FreeImage_GetBits(pBitmap) + nSrcPitch * (FreeImage_GetHeight(pBitmap) -1);
Npp8u *pDstLine = oImage.data();
unsigned int nDstPitch = oImage.pitch();
for (size_t iLine = 0; iLine < oImage.height(); ++iLine)
{
memcpy(pDstLine, pSrcLine, oImage.width() * sizeof(Npp8u));
pSrcLine -= nSrcPitch;
pDstLine += nDstPitch;
}
// swap the user given image with our result image, effecively
// moving our newly loaded image data into the user provided shell
oImage.swap(rImage);
}
// Save an gray-scale image to disk.
void
saveImage(const std::string &rFileName, const ImageCPU_8u_C1 &rImage)
{
// create the result image storage using FreeImage so we can easily
// save
FIBITMAP *pResultBitmap = FreeImage_Allocate(rImage.width(), rImage.height(), 8 /* bits per pixel */);
NPP_ASSERT_NOT_NULL(pResultBitmap);
unsigned int nDstPitch = FreeImage_GetPitch(pResultBitmap);
Npp8u *pDstLine = FreeImage_GetBits(pResultBitmap) + nDstPitch * (rImage.height()-1);
const Npp8u *pSrcLine = rImage.data();
unsigned int nSrcPitch = rImage.pitch();
for (size_t iLine = 0; iLine < rImage.height(); ++iLine)
{
memcpy(pDstLine, pSrcLine, rImage.width() * sizeof(Npp8u));
pSrcLine += nSrcPitch;
pDstLine -= nDstPitch;
}
// now save the result image
bool bSuccess;
bSuccess = FreeImage_Save(FIF_PGM, pResultBitmap, rFileName.c_str(), 0) == TRUE;
NPP_ASSERT_MSG(bSuccess, "Failed to save result image.");
}
// Load a gray-scale image from disk.
void
loadImage(const std::string &rFileName, ImageNPP_8u_C1 &rImage)
{
ImageCPU_8u_C1 oImage;
loadImage(rFileName, oImage);
ImageNPP_8u_C1 oResult(oImage);
rImage.swap(oResult);
}
// Save an gray-scale image to disk.
void
saveImage(const std::string &rFileName, const ImageNPP_8u_C1 &rImage)
{
ImageCPU_8u_C1 oHostImage(rImage.size());
// copy the device result data
rImage.copyTo(oHostImage.data(), oHostImage.pitch());
saveImage(rFileName, oHostImage);
}
}
#endif // NV_UTIL_NPP_IMAGE_IO_H

View File

@@ -0,0 +1,171 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGE_PACKED_H
#define NV_UTIL_NPP_IMAGE_PACKED_H
#include "Image.h"
#include "Pixel.h"
namespace npp
{
template<typename D, size_t N, class A>
class ImagePacked: public npp::Image
{
public:
typedef npp::Pixel<D, N> tPixel;
typedef D tData;
static const size_t gnChannels = N;
typedef npp::Image::Size tSize;
ImagePacked(): aPixels_(0)
, nPitch_(0)
{
;
}
ImagePacked(unsigned int nWidth, unsigned int nHeight): Image(nWidth, nHeight)
, aPixels_(0)
, nPitch_(0)
{
aPixels_ = A::Malloc2D(width(), height(), &nPitch_);
}
ImagePacked(unsigned int nWidth, unsigned int nHeight, bool bTight): Image(nWidth, nHeight)
, aPixels_(0)
, nPitch_(0)
{
aPixels_ = A::Malloc2D(width(), height(), &nPitch_, bTight);
}
ImagePacked(const tSize &rSize): Image(rSize)
, aPixels_(0)
, nPitch_(0)
{
aPixels_ = A::Malloc2D(width(), height(), &nPitch_);
}
ImagePacked(const ImagePacked<D, N, A> &rImage): Image(rImage)
, aPixels_(0)
, nPitch_(rImage.pitch())
{
aPixels_ = A::Malloc2D(width(), height(), &nPitch_);
A::Copy2D(aPixels_, nPitch_, rImage.pixels(), rImage.pitch(), width(), height());
}
virtual
~ImagePacked()
{
A::Free2D(aPixels_);
}
ImagePacked &
operator= (const ImagePacked<D, N, A> &rImage)
{
// in case of self-assignment
if (&rImage == this)
{
return *this;
}
A::Free2D(aPixels_);
aPixels_ = 0;
nPitch_ = 0;
// assign parent class's data fields (width, height)
Image::operator =(rImage);
aPixels_ = A::Malloc2D(width(), height(), &nPitch_);
A::Copy2D(aPixels_, nPitch_, rImage.data(), rImage.pitch(), width(), height());
return *this;
}
unsigned int
pitch()
const
{
return nPitch_;
}
/// Get a pointer to the pixel array.
/// The result pointer can be offset to pixel at position (x, y) and
/// even negative offsets are allowed.
/// \param nX Horizontal pointer/array offset.
/// \param nY Vertical pointer/array offset.
/// \return Pointer to the pixel array (or first pixel in array with coordinates (nX, nY).
tPixel *
pixels(int nX = 0, int nY = 0)
{
return reinterpret_cast<tPixel *>(reinterpret_cast<unsigned char *>(aPixels_) + nY * pitch() + nX * gnChannels * sizeof(D));
}
const
tPixel *
pixels(int nX = 0, int nY = 0)
const
{
return reinterpret_cast<const tPixel *>(reinterpret_cast<unsigned char *>(aPixels_) + nY * pitch() + nX * gnChannels * sizeof(D));
}
D *
data(int nX = 0, int nY = 0)
{
return reinterpret_cast<D *>(pixels(nX, nY));
}
const
D *
data(int nX = 0, int nY = 0)
const
{
return reinterpret_cast<const D *>(pixels(nX, nY));
}
void
swap(ImagePacked<D, N, A> &rImage)
{
Image::swap(rImage);
tData *aTemp = aPixels_;
aPixels_ = rImage.aPixels_;
rImage.aPixels_ = aTemp;
unsigned int nTemp = nPitch_;
nPitch_ = rImage.nPitch_;
rImage.nPitch_ = nTemp;
}
private:
D *aPixels_;
unsigned int nPitch_;
};
} // npp namespace
#endif // NV_IMAGE_IPP_H

121
Common/UtilNPP/ImagesCPU.h Normal file
View File

@@ -0,0 +1,121 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGES_CPU_H
#define NV_UTIL_NPP_IMAGES_CPU_H
#include "ImagePacked.h"
#include "ImageAllocatorsCPU.h"
#include "Exceptions.h"
#include <npp.h>
namespace npp
{
template<typename D, unsigned int N, class A>
class ImageCPU: public npp::ImagePacked<D, N, A>
{
public:
ImageCPU()
{
;
}
ImageCPU(unsigned int nWidth, unsigned int nHeight): ImagePacked<D, N, A>(nWidth, nHeight)
{
;
}
explicit
ImageCPU(const npp::Image::Size &rSize): ImagePacked<D, N, A>(rSize)
{
;
}
ImageCPU(const ImageCPU<D, N, A> &rImage): Image(rImage)
{
;
}
virtual
~ImageCPU()
{
;
}
ImageCPU &
operator= (const ImageCPU<D, N, A> &rImage)
{
ImagePacked<D, N, A>::operator= (rImage);
return *this;
}
npp::Pixel<D, N> &
operator()(unsigned int iX, unsigned int iY)
{
return *ImagePacked<D, N, A>::pixels(iX, iY);
}
npp::Pixel<D, N>
operator()(unsigned int iX, unsigned int iY)
const
{
return *ImagePacked<D, N, A>::pixels(iX, iY);
}
};
typedef ImageCPU<Npp8u, 1, npp::ImageAllocatorCPU<Npp8u, 1> > ImageCPU_8u_C1;
typedef ImageCPU<Npp8u, 2, npp::ImageAllocatorCPU<Npp8u, 2> > ImageCPU_8u_C2;
typedef ImageCPU<Npp8u, 3, npp::ImageAllocatorCPU<Npp8u, 3> > ImageCPU_8u_C3;
typedef ImageCPU<Npp8u, 4, npp::ImageAllocatorCPU<Npp8u, 4> > ImageCPU_8u_C4;
typedef ImageCPU<Npp16u, 1, npp::ImageAllocatorCPU<Npp16u, 1> > ImageCPU_16u_C1;
typedef ImageCPU<Npp16u, 3, npp::ImageAllocatorCPU<Npp16u, 3> > ImageCPU_16u_C3;
typedef ImageCPU<Npp16u, 4, npp::ImageAllocatorCPU<Npp16u, 4> > ImageCPU_16u_C4;
typedef ImageCPU<Npp16s, 1, npp::ImageAllocatorCPU<Npp16s, 1> > ImageCPU_16s_C1;
typedef ImageCPU<Npp16s, 3, npp::ImageAllocatorCPU<Npp16s, 3> > ImageCPU_16s_C3;
typedef ImageCPU<Npp16s, 4, npp::ImageAllocatorCPU<Npp16s, 4> > ImageCPU_16s_C4;
typedef ImageCPU<Npp32s, 1, npp::ImageAllocatorCPU<Npp32s, 1> > ImageCPU_32s_C1;
typedef ImageCPU<Npp32s, 3, npp::ImageAllocatorCPU<Npp32s, 3> > ImageCPU_32s_C3;
typedef ImageCPU<Npp32s, 4, npp::ImageAllocatorCPU<Npp32s, 4> > ImageCPU_32s_C4;
typedef ImageCPU<Npp32f, 1, npp::ImageAllocatorCPU<Npp32f, 1> > ImageCPU_32f_C1;
typedef ImageCPU<Npp32f, 3, npp::ImageAllocatorCPU<Npp32f, 3> > ImageCPU_32f_C3;
typedef ImageCPU<Npp32f, 4, npp::ImageAllocatorCPU<Npp32f, 4> > ImageCPU_32f_C4;
} // npp namespace
#endif // NV_IMAGE_IPP_H

149
Common/UtilNPP/ImagesNPP.h Normal file
View File

@@ -0,0 +1,149 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_IMAGES_NPP_H
#define NV_UTIL_NPP_IMAGES_NPP_H
#include "Exceptions.h"
#include "ImagePacked.h"
#include "ImageAllocatorsNPP.h"
#include <cuda_runtime.h>
namespace npp
{
// forward declaration
template<typename D, unsigned int N, class A> class ImageCPU;
template<typename D, unsigned int N>
class ImageNPP: public npp::ImagePacked<D, N, npp::ImageAllocator<D, N> >
{
public:
ImageNPP()
{
;
}
ImageNPP(unsigned int nWidth, unsigned int nHeight, bool bTight = false): ImagePacked<D, N, npp::ImageAllocator<D, N> >(nWidth, nHeight, bTight)
{
;
}
ImageNPP(const npp::Image::Size &rSize): ImagePacked<D, N, npp::ImageAllocator<D, N> >(rSize)
{
;
}
ImageNPP(const ImageNPP<D, N> &rImage): Image(rImage)
{
;
}
template<class X>
explicit
ImageNPP(const ImageCPU<D, N, X> &rImage, bool bTight = false): ImagePacked<D, N, npp::ImageAllocator<D, N> >(rImage.width(), rImage.height(), bTight)
{
npp::ImageAllocator<D, N>::HostToDeviceCopy2D(ImagePacked<D, N, npp::ImageAllocator<D, N> >::data(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::pitch(),
rImage.data(),
rImage.pitch(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::width(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::height());
}
virtual
~ImageNPP()
{
;
}
ImageNPP &
operator= (const ImageNPP<D, N> &rImage)
{
ImagePacked<D, N, npp::ImageAllocator<D, N> >::operator= (rImage);
return *this;
}
void
copyTo(D *pData, unsigned int nPitch)
const
{
NPP_ASSERT((ImagePacked<D, N, npp::ImageAllocator<D, N> >::width() * sizeof(npp::Pixel<D, N>) <= nPitch));
npp::ImageAllocator<D, N>::DeviceToHostCopy2D(pData,
nPitch,
ImagePacked<D, N, npp::ImageAllocator<D, N> >::data(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::pitch(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::width(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::height());
}
void
copyFrom(D *pData, unsigned int nPitch)
{
NPP_ASSERT((ImagePacked<D, N, npp::ImageAllocator<D, N> >::width() * sizeof(npp::Pixel<D, N>) <= nPitch));
npp::ImageAllocator<D, N>::HostToDeviceCopy2D(ImagePacked<D, N, npp::ImageAllocator<D, N> >::data(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::pitch(),
pData,
nPitch,
ImagePacked<D, N, npp::ImageAllocator<D, N> >::width(),
ImagePacked<D, N, npp::ImageAllocator<D, N> >::height());
}
};
typedef ImageNPP<Npp8u, 1> ImageNPP_8u_C1;
typedef ImageNPP<Npp8u, 2> ImageNPP_8u_C2;
typedef ImageNPP<Npp8u, 3> ImageNPP_8u_C3;
typedef ImageNPP<Npp8u, 4> ImageNPP_8u_C4;
typedef ImageNPP<Npp16u, 1> ImageNPP_16u_C1;
typedef ImageNPP<Npp16u, 2> ImageNPP_16u_C2;
typedef ImageNPP<Npp16u, 3> ImageNPP_16u_C3;
typedef ImageNPP<Npp16u, 4> ImageNPP_16u_C4;
typedef ImageNPP<Npp16s, 1> ImageNPP_16s_C1;
typedef ImageNPP<Npp16s, 3> ImageNPP_16s_C3;
typedef ImageNPP<Npp16s, 4> ImageNPP_16s_C4;
typedef ImageNPP<Npp32s, 1> ImageNPP_32s_C1;
typedef ImageNPP<Npp32s, 3> ImageNPP_32s_C3;
typedef ImageNPP<Npp32s, 4> ImageNPP_32s_C4;
typedef ImageNPP<Npp32f, 1> ImageNPP_32f_C1;
typedef ImageNPP<Npp32f, 2> ImageNPP_32f_C2;
typedef ImageNPP<Npp32f, 3> ImageNPP_32f_C3;
typedef ImageNPP<Npp32f, 4> ImageNPP_32f_C4;
typedef ImageNPP<Npp64f, 1> ImageNPP_64f_C1;
typedef ImageNPP<Npp64f, 2> ImageNPP_64f_C2;
typedef ImageNPP<Npp64f, 3> ImageNPP_64f_C3;
typedef ImageNPP<Npp64f, 4> ImageNPP_64f_C4;
} // npp namespace
#endif // NV_UTIL_NPP_IMAGES_NPP_H

126
Common/UtilNPP/Pixel.h Normal file
View File

@@ -0,0 +1,126 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_PIXEL_H
#define NV_UTIL_PIXEL_H
#include "Exceptions.h"
namespace npp
{
template <typename D, size_t N>
struct Pixel
{ };
template <typename D>
struct Pixel<D, 1>
{
D x;
const D &
operator[](size_t iChannel)
const
{
NPP_ASSERT(iChannel < 1);
return (&x)[iChannel];
}
D &
operator[](size_t iChannel)
{
NPP_ASSERT(iChannel < 1);
return (&x)[iChannel];
}
};
template <typename D>
struct Pixel<D, 2>
{
D x,y;
const D &
operator[](size_t iChannel)
const
{
NPP_ASSERT(iChannel < 2);
return (&x)[iChannel];
}
D &
operator[](size_t iChannel)
{
NPP_ASSERT(iChannel < 2);
return (&x)[iChannel];
}
};
template <typename D>
struct Pixel<D, 3>
{
D x,y,z;
const D &
operator[](size_t iChannel)
const
{
NPP_ASSERT(iChannel < 3);
return (&x)[iChannel];
}
D &
operator[](size_t iChannel)
{
NPP_ASSERT(iChannel < 3);
return (&x)[iChannel];
}
};
template <typename D>
struct Pixel<D, 4>
{
D x, y, z, w;
const D &
operator[](size_t iChannel)
const
{
NPP_ASSERT(iChannel < 4);
return (&x)[iChannel];
}
D &
operator[](size_t iChannel)
{
NPP_ASSERT(iChannel < 4);
return (&x)[iChannel];
}
};
} // npp namespace
#endif // NV_UTIL_PIXEL_H

168
Common/UtilNPP/Signal.h Normal file
View File

@@ -0,0 +1,168 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_SIGNAL_H
#define NV_UTIL_NPP_SIGNAL_H
#include <cstring>
namespace npp
{
class Signal
{
public:
Signal() : nSize_(0)
{ };
explicit
Signal(size_t nSize) : nSize_(nSize)
{ };
Signal(const Signal &rSignal) : nSize_(rSignal.nSize_)
{ };
virtual
~Signal()
{ }
Signal &
operator= (const Signal &rSignal)
{
nSize_ = rSignal.nSize_;
return *this;
}
size_t
size()
const
{
return nSize_;
}
void
swap(Signal &rSignal)
{
size_t nTemp = nSize_;
nSize_ = rSignal.nSize_;
rSignal.nSize_ = nTemp;
}
private:
size_t nSize_;
};
template<typename D, class A>
class SignalTemplate: public Signal
{
public:
typedef D tData;
SignalTemplate(): aValues_(0)
{
;
}
SignalTemplate(size_t nSize): Signal(nSize)
, aValues_(0)
{
aValues_ = A::Malloc1D(size());
}
SignalTemplate(const SignalTemplate<D, A> &rSignal): Signal(rSignal)
, aValues_(0)
{
aValues_ = A::Malloc1D(size());
A::Copy1D(aValues_, rSignal.values(), size());
}
virtual
~SignalTemplate()
{
A::Free1D(aValues_);
}
SignalTemplate &
operator= (const SignalTemplate<D, A> &rSignal)
{
// in case of self-assignment
if (&rSignal == this)
{
return *this;
}
A::Free1D(aValues_);
this->aPixels_ = 0;
// assign parent class's data fields (width, height)
Signal::operator =(rSignal);
aValues_ = A::Malloc1D(size());
A::Copy1D(aValues_, rSignal.value(), size());
return *this;
}
/// Get a pointer to the pixel array.
/// The result pointer can be offset to pixel at position (x, y) and
/// even negative offsets are allowed.
/// \param nX Horizontal pointer/array offset.
/// \param nY Vertical pointer/array offset.
/// \return Pointer to the pixel array (or first pixel in array with coordinates (nX, nY).
tData *
values(int i = 0)
{
return aValues_ + i;
}
const
tData *
values(int i = 0)
const
{
return aValues_ + i;
}
void
swap(SignalTemplate<D, A> &rSignal)
{
Signal::swap(rSignal);
tData *aTemp = this->aValues_;
this->aValues_ = rSignal.aValues_;
rSignal.aValues_ = aTemp;
}
private:
D *aValues_;
};
} // npp namespace
#endif // NV_UTIL_NPP_SIGNAL_H

View File

@@ -0,0 +1,66 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_SIGNAL_ALLOCATORS_CPU_H
#define NV_UTIL_NPP_SIGNAL_ALLOCATORS_CPU_H
#include "Exceptions.h"
namespace npp
{
template <typename D>
class SignalAllocatorCPU
{
public:
static
D *
Malloc1D(unsigned int nSize)
{
return new D[nSize];;
};
static
void
Free1D(D *pPixels)
{
delete[] pPixels;
};
static
void
Copy1D(D *pDst, const D *pSrc, size_t nSize)
{
memcpy(pDst, pSrc, nSize * sizeof(D));
};
};
} // npp namespace
#endif // NV_UTIL_NPP_SIGNAL_ALLOCATORS_CPU_H

View File

@@ -0,0 +1,684 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_SIGNAL_ALLOCATORS_NPP_H
#define NV_UTIL_NPP_SIGNAL_ALLOCATORS_NPP_H
#include "Exceptions.h"
#include <npps.h>
#include <cuda_runtime.h>
namespace npp
{
template <typename D>
class SignalAllocator
{
};
template<>
class SignalAllocator<Npp8u>
{
public:
static
Npp8u *
Malloc1D(size_t nSize)
{
Npp8u *pResult = nppsMalloc_8u(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp8u *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp8u *pDst, const Npp8u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp8u),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp8u *pDst, const Npp8u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp8u), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp8u *pDst, const Npp8u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp8u), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp16s>
{
public:
static
Npp16s *
Malloc1D(size_t nSize)
{
Npp16s *pResult = nppsMalloc_16s(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp16s *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp16s *pDst, const Npp16s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16s),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp16s *pDst, const Npp16s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16s), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp16s *pDst, const Npp16s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16s), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp16u>
{
public:
static
Npp16u *
Malloc1D(size_t nSize)
{
Npp16u *pResult = nppsMalloc_16u(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp16u *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp16u *pDst, const Npp16u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16u),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp16u *pDst, const Npp16u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16u), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp16u *pDst, const Npp16u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16u), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp16sc>
{
public:
static
Npp16sc *
Malloc1D(size_t nSize)
{
Npp16sc *pResult = nppsMalloc_16sc(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp16sc *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp16sc *pDst, const Npp16sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16sc),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp16sc *pDst, const Npp16sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16sc), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp16sc *pDst, const Npp16sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp16sc), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp32u>
{
public:
static
Npp32u *
Malloc1D(size_t nSize)
{
Npp32u *pResult = nppsMalloc_32u(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp32u *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp32u *pDst, const Npp32u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32u),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp32u *pDst, const Npp32u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32u), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp32u *pDst, const Npp32u *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32u), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp32s>
{
public:
static
Npp32s *
Malloc1D(size_t nSize)
{
Npp32s *pResult = nppsMalloc_32s(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp32s *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp32s *pDst, const Npp32s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32s),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp32s *pDst, const Npp32s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32s), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp32s *pDst, const Npp32s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32s), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp32sc>
{
public:
static
Npp32sc *
Malloc1D(size_t nSize)
{
Npp32sc *pResult = nppsMalloc_32sc(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp32sc *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp32sc *pDst, const Npp32sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32sc),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp32sc *pDst, const Npp32sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32sc), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp32sc *pDst, const Npp32sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32sc), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp32f>
{
public:
static
Npp32f *
Malloc1D(size_t nSize)
{
Npp32f *pResult = nppsMalloc_32f(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp32f *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp32f *pDst, const Npp32f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32f),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp32f *pDst, const Npp32f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32f), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp32f *pDst, const Npp32f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32f), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp32fc>
{
public:
static
Npp32fc *
Malloc1D(size_t nSize)
{
Npp32fc *pResult = nppsMalloc_32fc(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp32fc *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp32fc *pDst, const Npp32fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32fc),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp32fc *pDst, const Npp32fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32fc), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp32fc *pDst, const Npp32fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp32fc), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp64s>
{
public:
static
Npp64s *
Malloc1D(size_t nSize)
{
Npp64s *pResult = nppsMalloc_64s(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp64s *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp64s *pDst, const Npp64s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64s),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp64s *pDst, const Npp64s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64s), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp64s *pDst, const Npp64s *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64s), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp64sc>
{
public:
static
Npp64sc *
Malloc1D(size_t nSize)
{
Npp64sc *pResult = nppsMalloc_64sc(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp64sc *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp64sc *pDst, const Npp64sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64sc),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp64sc *pDst, const Npp64sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64sc), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp64sc *pDst, const Npp64sc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64sc), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp64f>
{
public:
static
Npp64f *
Malloc1D(size_t nSize)
{
Npp64f *pResult = nppsMalloc_64f(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp64f *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp64f *pDst, const Npp64f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64f),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp64f *pDst, const Npp64f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64f), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp64f *pDst, const Npp64f *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64f), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
template<>
class SignalAllocator<Npp64fc>
{
public:
static
Npp64fc *
Malloc1D(size_t nSize)
{
Npp64fc *pResult = nppsMalloc_64fc(static_cast<int>(nSize));
NPP_ASSERT(pResult != 0);
return pResult;
};
static
void
Free1D(Npp64fc *pValues)
{
nppsFree(pValues);
};
static
void
Copy1D(Npp64fc *pDst, const Npp64fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64fc),cudaMemcpyDeviceToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
HostToDeviceCopy1D(Npp64fc *pDst, const Npp64fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64fc), cudaMemcpyHostToDevice);
NPP_ASSERT(cudaSuccess == eResult);
};
static
void
DeviceToHostCopy1D(Npp64fc *pDst, const Npp64fc *pSrc, size_t nSize)
{
cudaError_t eResult;
eResult = cudaMemcpy(pDst, pSrc, nSize * sizeof(Npp64fc), cudaMemcpyDeviceToHost);
NPP_ASSERT(cudaSuccess == eResult);
};
};
} // npp namespace
#endif // NV_UTIL_NPP_SIGNAL_ALLOCATORS_NPP_H

107
Common/UtilNPP/SignalsCPU.h Normal file
View File

@@ -0,0 +1,107 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_SIGNALS_CPU_H
#define NV_UTIL_NPP_SIGNALS_CPU_H
#include "Signal.h"
#include "SignalAllocatorsCPU.h"
#include "Exceptions.h"
#include <npp.h>
namespace npp
{
template<typename D, class A>
class SignalCPU: public npp::SignalTemplate<D, A>
{
public:
typedef typename npp::SignalTemplate<D, A>::tData tData;
SignalCPU()
{
;
}
SignalCPU(size_t nSize): SignalTemplate<D, A>(nSize)
{
;
}
SignalCPU(const SignalCPU<D, A> &rSignal): SignalTemplate<D, A>(rSignal)
{
;
}
virtual
~SignalCPU()
{
;
}
SignalCPU &
operator= (const SignalCPU<D,A> &rSignal)
{
SignalTemplate<D, A>::operator= (rSignal);
return *this;
}
tData &
operator [](unsigned int i)
{
return *SignalTemplate<D, A>::values(i);
}
tData
operator [](unsigned int i)
const
{
return *SignalTemplate<D, A>::values(i);
}
};
typedef SignalCPU<Npp8u, npp::SignalAllocatorCPU<Npp8u> > SignalCPU_8u;
typedef SignalCPU<Npp32s, npp::SignalAllocatorCPU<Npp32s> > SignalCPU_32s;
typedef SignalCPU<Npp16s, npp::SignalAllocatorCPU<Npp16s> > SignalCPU_16s;
typedef SignalCPU<Npp16sc, npp::SignalAllocatorCPU<Npp16sc> > SignalCPU_16sc;
typedef SignalCPU<Npp32sc, npp::SignalAllocatorCPU<Npp32sc> > SignalCPU_32sc;
typedef SignalCPU<Npp32f, npp::SignalAllocatorCPU<Npp32f> > SignalCPU_32f;
typedef SignalCPU<Npp32fc, npp::SignalAllocatorCPU<Npp32fc> > SignalCPU_32fc;
typedef SignalCPU<Npp64s, npp::SignalAllocatorCPU<Npp64s> > SignalCPU_64s;
typedef SignalCPU<Npp64sc, npp::SignalAllocatorCPU<Npp64sc> > SignalCPU_64sc;
typedef SignalCPU<Npp64f, npp::SignalAllocatorCPU<Npp64f> > SignalCPU_64f;
typedef SignalCPU<Npp64fc, npp::SignalAllocatorCPU<Npp64fc> > SignalCPU_64fc;
} // npp namespace
#endif // NV_UTIL_NPP_SIGNALS_CPU_H

113
Common/UtilNPP/SignalsNPP.h Normal file
View File

@@ -0,0 +1,113 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef NV_UTIL_NPP_SIGNALS_NPP_H
#define NV_UTIL_NPP_SIGNALS_NPP_H
#include "Exceptions.h"
#include "Signal.h"
#include "SignalAllocatorsNPP.h"
#include <cuda_runtime.h>
namespace npp
{
// forward declaration
template<typename D, class A> class SignalCPU;
template<typename D>
class SignalNPP: public npp::SignalTemplate<D, npp::SignalAllocator<D> >
{
public:
SignalNPP()
{
;
}
explicit
SignalNPP(size_t nSize): SignalTemplate<D, npp::SignalAllocator<D> >(nSize)
{
;
}
SignalNPP(const SignalNPP<D> &rSignal): SignalTemplate<D, npp::SignalAllocator<D> >(rSignal)
{
;
}
template<class X>
explicit
SignalNPP(const SignalCPU<D, X> &rSignal): SignalTemplate<D, npp::SignalAllocator<D> >(rSignal.size())
{
npp::SignalAllocator<D>::HostToDeviceCopy1D(SignalTemplate<D, npp::SignalAllocator<D> >::values(),
rSignal.values(), SignalTemplate<D, npp::SignalAllocator<D> >::size());
}
virtual
~SignalNPP()
{
;
}
SignalNPP &
operator= (const SignalNPP<D> &rSignal)
{
SignalTemplate<D, npp::SignalAllocator<D> >::operator= (rSignal);
return *this;
}
void
copyTo(D *pValues)
const
{
npp::SignalAllocator<D>::DeviceToHostCopy1D(pValues, SignalTemplate<D, npp::SignalAllocator<D> >::values(), SignalTemplate<D, npp::SignalAllocator<D> >::size());
}
void
copyFrom(D *pValues)
{
npp::SignalAllocator<D>::HostToDeviceCopy1D(SignalTemplate<D, npp::SignalAllocator<D> >::values(), pValues, SignalTemplate<D, npp::SignalAllocator<D> >::size());
}
};
typedef SignalNPP<Npp8u> SignalNPP_8u;
typedef SignalNPP<Npp16s> SignalNPP_16s;
typedef SignalNPP<Npp16sc> SignalNPP_16sc;
typedef SignalNPP<Npp32s> SignalNPP_32s;
typedef SignalNPP<Npp32sc> SignalNPP_32sc;
typedef SignalNPP<Npp32f> SignalNPP_32f;
typedef SignalNPP<Npp32fc> SignalNPP_32fc;
typedef SignalNPP<Npp64s> SignalNPP_64s;
typedef SignalNPP<Npp64sc> SignalNPP_64sc;
typedef SignalNPP<Npp64f> SignalNPP_64f;
typedef SignalNPP<Npp64fc> SignalNPP_64fc;
} // npp namespace
#endif // NV_UTIL_NPP_SIGNALS_NPP_H

Binary file not shown.

Binary file not shown.

File diff suppressed because one or more lines are too long

Binary file not shown.

Binary file not shown.

Binary file not shown.

File diff suppressed because one or more lines are too long

Binary file not shown.

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

160
Common/dynlink_d3d11.h Normal file
View File

@@ -0,0 +1,160 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//--------------------------------------------------------------------------------------
// File: dynlink_d3d11.h
//
// Shortcut macros and functions for using DX objects
//
// Copyright (c) Microsoft Corporation. All rights reserved
//--------------------------------------------------------------------------------------
#ifndef _DYNLINK_D3D11_H_
#define _DYNLINK_D3D11_H_
// Standard Windows includes
#include <windows.h>
#include <initguid.h>
#include <assert.h>
#include <wchar.h>
#include <mmsystem.h>
#include <commctrl.h> // for InitCommonControls()
#include <shellapi.h> // for ExtractIcon()
#include <new.h> // for placement new
#include <shlobj.h>
#include <math.h>
#include <limits.h>
#include <stdio.h>
// CRT's memory leak detection
#if defined(DEBUG) || defined(_DEBUG)
#include <crtdbg.h>
#endif
// Direct3D10 includes
#include <dxgi.h>
#include <d3d11.h>
// #include <..\Samples\C++\Effects11\Inc\d3dx11effect.h>
// XInput includes
#include <xinput.h>
// strsafe.h deprecates old unsecure string functions. If you
// really do not want to it to (not recommended), then uncomment the next line
//#define STRSAFE_NO_DEPRECATE
#ifndef STRSAFE_NO_DEPRECATE
#pragma deprecated("strncpy")
#pragma deprecated("wcsncpy")
#pragma deprecated("_tcsncpy")
#pragma deprecated("wcsncat")
#pragma deprecated("strncat")
#pragma deprecated("_tcsncat")
#endif
#pragma warning( disable : 4996 ) // disable deprecated warning
#include <strsafe.h>
#pragma warning( default : 4996 )
typedef HRESULT(WINAPI *LPCREATEDXGIFACTORY)(REFIID, void **);
typedef HRESULT(WINAPI *LPD3D11CREATEDEVICEANDSWAPCHAIN)(__in_opt IDXGIAdapter *pAdapter, D3D_DRIVER_TYPE DriverType, HMODULE Software, UINT Flags, __in_ecount_opt(FeatureLevels) CONST D3D_FEATURE_LEVEL *pFeatureLevels, UINT FeatureLevels, UINT SDKVersion, __in_opt CONST DXGI_SWAP_CHAIN_DESC *pSwapChainDesc, __out_opt IDXGISwapChain **ppSwapChain, __out_opt ID3D11Device **ppDevice, __out_opt D3D_FEATURE_LEVEL *pFeatureLevel, __out_opt ID3D11DeviceContext **ppImmediateContext);
typedef HRESULT(WINAPI *LPD3D11CREATEDEVICE)(IDXGIAdapter *, D3D_DRIVER_TYPE, HMODULE, UINT32, D3D_FEATURE_LEVEL *, UINT, UINT32, ID3D11Device **, D3D_FEATURE_LEVEL *, ID3D11DeviceContext **);
static HMODULE s_hModDXGI = NULL;
static LPCREATEDXGIFACTORY sFnPtr_CreateDXGIFactory = NULL;
static HMODULE s_hModD3D11 = NULL;
static LPD3D11CREATEDEVICE sFnPtr_D3D11CreateDevice = NULL;
static LPD3D11CREATEDEVICEANDSWAPCHAIN sFnPtr_D3D11CreateDeviceAndSwapChain = NULL;
// unload the D3D10 DLLs
static bool dynlinkUnloadD3D11API(void)
{
if (s_hModDXGI)
{
FreeLibrary(s_hModDXGI);
s_hModDXGI = NULL;
}
if (s_hModD3D11)
{
FreeLibrary(s_hModD3D11);
s_hModD3D11 = NULL;
}
return true;
}
// Dynamically load the D3D11 DLLs loaded and map the function pointers
static bool dynlinkLoadD3D11API(void)
{
// If both modules are non-NULL, this function has already been called. Note
// that this doesn't guarantee that all ProcAddresses were found.
if (s_hModD3D11 != NULL && s_hModDXGI != NULL)
{
return true;
}
#if 1
// This may fail if Direct3D 11 isn't installed
s_hModD3D11 = LoadLibrary("d3d11.dll");
if (s_hModD3D11 != NULL)
{
sFnPtr_D3D11CreateDevice = (LPD3D11CREATEDEVICE)GetProcAddress(s_hModD3D11, "D3D11CreateDevice");
sFnPtr_D3D11CreateDeviceAndSwapChain = (LPD3D11CREATEDEVICEANDSWAPCHAIN)GetProcAddress(s_hModD3D11, "D3D11CreateDeviceAndSwapChain");
}
else
{
printf("\nLoad d3d11.dll failed\n");
fflush(0);
}
if (!sFnPtr_CreateDXGIFactory)
{
s_hModDXGI = LoadLibrary("dxgi.dll");
if (s_hModDXGI)
{
sFnPtr_CreateDXGIFactory = (LPCREATEDXGIFACTORY)GetProcAddress(s_hModDXGI, "CreateDXGIFactory1");
}
return (s_hModDXGI != NULL) && (s_hModD3D11 != NULL);
}
return (s_hModD3D11 != NULL);
#else
sFnPtr_D3D11CreateDevice = (LPD3D11CREATEDEVICE)D3D11CreateDeviceAndSwapChain;
sFnPtr_D3D11CreateDeviceAndSwapChain = (LPD3D11CREATEDEVICEANDSWAPCHAIN)D3D11CreateDeviceAndSwapChain;
//sFnPtr_D3DX11CreateEffectFromMemory = ( LPD3DX11CREATEEFFECTFROMMEMORY )D3DX11CreateEffectFromMemory;
sFnPtr_D3DX11CompileFromMemory = (LPD3DX11COMPILEFROMMEMORY)D3DX11CompileFromMemory;
sFnPtr_CreateDXGIFactory = (LPCREATEDXGIFACTORY)CreateDXGIFactory;
return true;
#endif
return true;
}
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -51,457 +51,17 @@
// CUDA Runtime error messages // CUDA Runtime error messages
#ifdef __DRIVER_TYPES_H__ #ifdef __DRIVER_TYPES_H__
static const char *_cudaGetErrorEnum(cudaError_t error) { static const char *_cudaGetErrorEnum(cudaError_t error) {
switch (error) { return cudaGetErrorName(error);
case cudaSuccess:
return "cudaSuccess";
case cudaErrorMissingConfiguration:
return "cudaErrorMissingConfiguration";
case cudaErrorMemoryAllocation:
return "cudaErrorMemoryAllocation";
case cudaErrorInitializationError:
return "cudaErrorInitializationError";
case cudaErrorLaunchFailure:
return "cudaErrorLaunchFailure";
case cudaErrorPriorLaunchFailure:
return "cudaErrorPriorLaunchFailure";
case cudaErrorLaunchTimeout:
return "cudaErrorLaunchTimeout";
case cudaErrorLaunchOutOfResources:
return "cudaErrorLaunchOutOfResources";
case cudaErrorInvalidDeviceFunction:
return "cudaErrorInvalidDeviceFunction";
case cudaErrorInvalidConfiguration:
return "cudaErrorInvalidConfiguration";
case cudaErrorInvalidDevice:
return "cudaErrorInvalidDevice";
case cudaErrorInvalidValue:
return "cudaErrorInvalidValue";
case cudaErrorInvalidPitchValue:
return "cudaErrorInvalidPitchValue";
case cudaErrorInvalidSymbol:
return "cudaErrorInvalidSymbol";
case cudaErrorMapBufferObjectFailed:
return "cudaErrorMapBufferObjectFailed";
case cudaErrorUnmapBufferObjectFailed:
return "cudaErrorUnmapBufferObjectFailed";
case cudaErrorInvalidHostPointer:
return "cudaErrorInvalidHostPointer";
case cudaErrorInvalidDevicePointer:
return "cudaErrorInvalidDevicePointer";
case cudaErrorInvalidTexture:
return "cudaErrorInvalidTexture";
case cudaErrorInvalidTextureBinding:
return "cudaErrorInvalidTextureBinding";
case cudaErrorInvalidChannelDescriptor:
return "cudaErrorInvalidChannelDescriptor";
case cudaErrorInvalidMemcpyDirection:
return "cudaErrorInvalidMemcpyDirection";
case cudaErrorAddressOfConstant:
return "cudaErrorAddressOfConstant";
case cudaErrorTextureFetchFailed:
return "cudaErrorTextureFetchFailed";
case cudaErrorTextureNotBound:
return "cudaErrorTextureNotBound";
case cudaErrorSynchronizationError:
return "cudaErrorSynchronizationError";
case cudaErrorInvalidFilterSetting:
return "cudaErrorInvalidFilterSetting";
case cudaErrorInvalidNormSetting:
return "cudaErrorInvalidNormSetting";
case cudaErrorMixedDeviceExecution:
return "cudaErrorMixedDeviceExecution";
case cudaErrorCudartUnloading:
return "cudaErrorCudartUnloading";
case cudaErrorUnknown:
return "cudaErrorUnknown";
case cudaErrorNotYetImplemented:
return "cudaErrorNotYetImplemented";
case cudaErrorMemoryValueTooLarge:
return "cudaErrorMemoryValueTooLarge";
case cudaErrorInvalidResourceHandle:
return "cudaErrorInvalidResourceHandle";
case cudaErrorNotReady:
return "cudaErrorNotReady";
case cudaErrorInsufficientDriver:
return "cudaErrorInsufficientDriver";
case cudaErrorSetOnActiveProcess:
return "cudaErrorSetOnActiveProcess";
case cudaErrorInvalidSurface:
return "cudaErrorInvalidSurface";
case cudaErrorNoDevice:
return "cudaErrorNoDevice";
case cudaErrorECCUncorrectable:
return "cudaErrorECCUncorrectable";
case cudaErrorSharedObjectSymbolNotFound:
return "cudaErrorSharedObjectSymbolNotFound";
case cudaErrorSharedObjectInitFailed:
return "cudaErrorSharedObjectInitFailed";
case cudaErrorUnsupportedLimit:
return "cudaErrorUnsupportedLimit";
case cudaErrorDuplicateVariableName:
return "cudaErrorDuplicateVariableName";
case cudaErrorDuplicateTextureName:
return "cudaErrorDuplicateTextureName";
case cudaErrorDuplicateSurfaceName:
return "cudaErrorDuplicateSurfaceName";
case cudaErrorDevicesUnavailable:
return "cudaErrorDevicesUnavailable";
case cudaErrorInvalidKernelImage:
return "cudaErrorInvalidKernelImage";
case cudaErrorNoKernelImageForDevice:
return "cudaErrorNoKernelImageForDevice";
case cudaErrorIncompatibleDriverContext:
return "cudaErrorIncompatibleDriverContext";
case cudaErrorPeerAccessAlreadyEnabled:
return "cudaErrorPeerAccessAlreadyEnabled";
case cudaErrorPeerAccessNotEnabled:
return "cudaErrorPeerAccessNotEnabled";
case cudaErrorDeviceAlreadyInUse:
return "cudaErrorDeviceAlreadyInUse";
case cudaErrorProfilerDisabled:
return "cudaErrorProfilerDisabled";
case cudaErrorProfilerNotInitialized:
return "cudaErrorProfilerNotInitialized";
case cudaErrorProfilerAlreadyStarted:
return "cudaErrorProfilerAlreadyStarted";
case cudaErrorProfilerAlreadyStopped:
return "cudaErrorProfilerAlreadyStopped";
/* Since CUDA 4.0*/
case cudaErrorAssert:
return "cudaErrorAssert";
case cudaErrorTooManyPeers:
return "cudaErrorTooManyPeers";
case cudaErrorHostMemoryAlreadyRegistered:
return "cudaErrorHostMemoryAlreadyRegistered";
case cudaErrorHostMemoryNotRegistered:
return "cudaErrorHostMemoryNotRegistered";
/* Since CUDA 5.0 */
case cudaErrorOperatingSystem:
return "cudaErrorOperatingSystem";
case cudaErrorPeerAccessUnsupported:
return "cudaErrorPeerAccessUnsupported";
case cudaErrorLaunchMaxDepthExceeded:
return "cudaErrorLaunchMaxDepthExceeded";
case cudaErrorLaunchFileScopedTex:
return "cudaErrorLaunchFileScopedTex";
case cudaErrorLaunchFileScopedSurf:
return "cudaErrorLaunchFileScopedSurf";
case cudaErrorSyncDepthExceeded:
return "cudaErrorSyncDepthExceeded";
case cudaErrorLaunchPendingCountExceeded:
return "cudaErrorLaunchPendingCountExceeded";
case cudaErrorNotPermitted:
return "cudaErrorNotPermitted";
case cudaErrorNotSupported:
return "cudaErrorNotSupported";
/* Since CUDA 6.0 */
case cudaErrorHardwareStackError:
return "cudaErrorHardwareStackError";
case cudaErrorIllegalInstruction:
return "cudaErrorIllegalInstruction";
case cudaErrorMisalignedAddress:
return "cudaErrorMisalignedAddress";
case cudaErrorInvalidAddressSpace:
return "cudaErrorInvalidAddressSpace";
case cudaErrorInvalidPc:
return "cudaErrorInvalidPc";
case cudaErrorIllegalAddress:
return "cudaErrorIllegalAddress";
/* Since CUDA 6.5*/
case cudaErrorInvalidPtx:
return "cudaErrorInvalidPtx";
case cudaErrorInvalidGraphicsContext:
return "cudaErrorInvalidGraphicsContext";
case cudaErrorStartupFailure:
return "cudaErrorStartupFailure";
case cudaErrorApiFailureBase:
return "cudaErrorApiFailureBase";
/* Since CUDA 8.0*/
case cudaErrorNvlinkUncorrectable:
return "cudaErrorNvlinkUncorrectable";
/* Since CUDA 8.5*/
case cudaErrorJitCompilerNotFound:
return "cudaErrorJitCompilerNotFound";
/* Since CUDA 9.0*/
case cudaErrorCooperativeLaunchTooLarge:
return "cudaErrorCooperativeLaunchTooLarge";
}
return "<unknown>";
} }
#endif #endif
#ifdef __cuda_cuda_h__ #ifdef CUDA_DRIVER_API
// CUDA Driver API errors // CUDA Driver API errors
static const char *_cudaGetErrorEnum(CUresult error) { static const char *_cudaGetErrorEnum(CUresult error) {
switch (error) { static char unknown[] = "<unknown>";
case CUDA_SUCCESS: const char *ret = NULL;
return "CUDA_SUCCESS"; cuGetErrorName(error, &ret);
return ret ? ret : unknown;
case CUDA_ERROR_INVALID_VALUE:
return "CUDA_ERROR_INVALID_VALUE";
case CUDA_ERROR_OUT_OF_MEMORY:
return "CUDA_ERROR_OUT_OF_MEMORY";
case CUDA_ERROR_NOT_INITIALIZED:
return "CUDA_ERROR_NOT_INITIALIZED";
case CUDA_ERROR_DEINITIALIZED:
return "CUDA_ERROR_DEINITIALIZED";
case CUDA_ERROR_PROFILER_DISABLED:
return "CUDA_ERROR_PROFILER_DISABLED";
case CUDA_ERROR_PROFILER_NOT_INITIALIZED:
return "CUDA_ERROR_PROFILER_NOT_INITIALIZED";
case CUDA_ERROR_PROFILER_ALREADY_STARTED:
return "CUDA_ERROR_PROFILER_ALREADY_STARTED";
case CUDA_ERROR_PROFILER_ALREADY_STOPPED:
return "CUDA_ERROR_PROFILER_ALREADY_STOPPED";
case CUDA_ERROR_NO_DEVICE:
return "CUDA_ERROR_NO_DEVICE";
case CUDA_ERROR_INVALID_DEVICE:
return "CUDA_ERROR_INVALID_DEVICE";
case CUDA_ERROR_INVALID_IMAGE:
return "CUDA_ERROR_INVALID_IMAGE";
case CUDA_ERROR_INVALID_CONTEXT:
return "CUDA_ERROR_INVALID_CONTEXT";
case CUDA_ERROR_CONTEXT_ALREADY_CURRENT:
return "CUDA_ERROR_CONTEXT_ALREADY_CURRENT";
case CUDA_ERROR_MAP_FAILED:
return "CUDA_ERROR_MAP_FAILED";
case CUDA_ERROR_UNMAP_FAILED:
return "CUDA_ERROR_UNMAP_FAILED";
case CUDA_ERROR_ARRAY_IS_MAPPED:
return "CUDA_ERROR_ARRAY_IS_MAPPED";
case CUDA_ERROR_ALREADY_MAPPED:
return "CUDA_ERROR_ALREADY_MAPPED";
case CUDA_ERROR_NO_BINARY_FOR_GPU:
return "CUDA_ERROR_NO_BINARY_FOR_GPU";
case CUDA_ERROR_ALREADY_ACQUIRED:
return "CUDA_ERROR_ALREADY_ACQUIRED";
case CUDA_ERROR_NOT_MAPPED:
return "CUDA_ERROR_NOT_MAPPED";
case CUDA_ERROR_NOT_MAPPED_AS_ARRAY:
return "CUDA_ERROR_NOT_MAPPED_AS_ARRAY";
case CUDA_ERROR_NOT_MAPPED_AS_POINTER:
return "CUDA_ERROR_NOT_MAPPED_AS_POINTER";
case CUDA_ERROR_ECC_UNCORRECTABLE:
return "CUDA_ERROR_ECC_UNCORRECTABLE";
case CUDA_ERROR_UNSUPPORTED_LIMIT:
return "CUDA_ERROR_UNSUPPORTED_LIMIT";
case CUDA_ERROR_CONTEXT_ALREADY_IN_USE:
return "CUDA_ERROR_CONTEXT_ALREADY_IN_USE";
case CUDA_ERROR_PEER_ACCESS_UNSUPPORTED:
return "CUDA_ERROR_PEER_ACCESS_UNSUPPORTED";
case CUDA_ERROR_INVALID_PTX:
return "CUDA_ERROR_INVALID_PTX";
case CUDA_ERROR_INVALID_GRAPHICS_CONTEXT:
return "CUDA_ERROR_INVALID_GRAPHICS_CONTEXT";
case CUDA_ERROR_NVLINK_UNCORRECTABLE:
return "CUDA_ERROR_NVLINK_UNCORRECTABLE";
case CUDA_ERROR_JIT_COMPILER_NOT_FOUND:
return "CUDA_ERROR_JIT_COMPILER_NOT_FOUND";
case CUDA_ERROR_INVALID_SOURCE:
return "CUDA_ERROR_INVALID_SOURCE";
case CUDA_ERROR_FILE_NOT_FOUND:
return "CUDA_ERROR_FILE_NOT_FOUND";
case CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND:
return "CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND";
case CUDA_ERROR_SHARED_OBJECT_INIT_FAILED:
return "CUDA_ERROR_SHARED_OBJECT_INIT_FAILED";
case CUDA_ERROR_OPERATING_SYSTEM:
return "CUDA_ERROR_OPERATING_SYSTEM";
case CUDA_ERROR_INVALID_HANDLE:
return "CUDA_ERROR_INVALID_HANDLE";
case CUDA_ERROR_NOT_FOUND:
return "CUDA_ERROR_NOT_FOUND";
case CUDA_ERROR_NOT_READY:
return "CUDA_ERROR_NOT_READY";
case CUDA_ERROR_ILLEGAL_ADDRESS:
return "CUDA_ERROR_ILLEGAL_ADDRESS";
case CUDA_ERROR_LAUNCH_FAILED:
return "CUDA_ERROR_LAUNCH_FAILED";
case CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES:
return "CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES";
case CUDA_ERROR_LAUNCH_TIMEOUT:
return "CUDA_ERROR_LAUNCH_TIMEOUT";
case CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING:
return "CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING";
case CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED:
return "CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED";
case CUDA_ERROR_PEER_ACCESS_NOT_ENABLED:
return "CUDA_ERROR_PEER_ACCESS_NOT_ENABLED";
case CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE:
return "CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE";
case CUDA_ERROR_CONTEXT_IS_DESTROYED:
return "CUDA_ERROR_CONTEXT_IS_DESTROYED";
case CUDA_ERROR_ASSERT:
return "CUDA_ERROR_ASSERT";
case CUDA_ERROR_TOO_MANY_PEERS:
return "CUDA_ERROR_TOO_MANY_PEERS";
case CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED:
return "CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED";
case CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED:
return "CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED";
case CUDA_ERROR_HARDWARE_STACK_ERROR:
return "CUDA_ERROR_HARDWARE_STACK_ERROR";
case CUDA_ERROR_ILLEGAL_INSTRUCTION:
return "CUDA_ERROR_ILLEGAL_INSTRUCTION";
case CUDA_ERROR_MISALIGNED_ADDRESS:
return "CUDA_ERROR_MISALIGNED_ADDRESS";
case CUDA_ERROR_INVALID_ADDRESS_SPACE:
return "CUDA_ERROR_INVALID_ADDRESS_SPACE";
case CUDA_ERROR_INVALID_PC:
return "CUDA_ERROR_INVALID_PC";
case CUDA_ERROR_COOPERATIVE_LAUNCH_TOO_LARGE:
return "CUDA_ERROR_COOPERATIVE_LAUNCH_TOO_LARGE";
case CUDA_ERROR_NOT_PERMITTED:
return "CUDA_ERROR_NOT_PERMITTED";
case CUDA_ERROR_NOT_SUPPORTED:
return "CUDA_ERROR_NOT_SUPPORTED";
case CUDA_ERROR_UNKNOWN:
return "CUDA_ERROR_UNKNOWN";
}
return "<unknown>";
} }
#endif #endif
@@ -722,6 +282,42 @@ static const char *_cudaGetErrorEnum(curandStatus_t error) {
} }
#endif #endif
#ifdef NVJPEGAPI
// nvJPEG API errors
static const char *_cudaGetErrorEnum(nvjpegStatus_t error) {
switch (error) {
case NVJPEG_STATUS_SUCCESS:
return "NVJPEG_STATUS_SUCCESS";
case NVJPEG_STATUS_NOT_INITIALIZED:
return "NVJPEG_STATUS_NOT_INITIALIZED";
case NVJPEG_STATUS_INVALID_PARAMETER:
return "NVJPEG_STATUS_INVALID_PARAMETER";
case NVJPEG_STATUS_BAD_JPEG:
return "NVJPEG_STATUS_BAD_JPEG";
case NVJPEG_STATUS_JPEG_NOT_SUPPORTED:
return "NVJPEG_STATUS_JPEG_NOT_SUPPORTED";
case NVJPEG_STATUS_ALLOCATOR_FAILURE:
return "NVJPEG_STATUS_ALLOCATOR_FAILURE";
case NVJPEG_STATUS_EXECUTION_FAILED:
return "NVJPEG_STATUS_EXECUTION_FAILED";
case NVJPEG_STATUS_ARCH_MISMATCH:
return "NVJPEG_STATUS_ARCH_MISMATCH";
case NVJPEG_STATUS_INTERNAL_ERROR:
return "NVJPEG_STATUS_INTERNAL_ERROR";
}
return "<unknown>";
}
#endif
#ifdef NV_NPPIDEFS_H #ifdef NV_NPPIDEFS_H
// NPP API errors // NPP API errors
static const char *_cudaGetErrorEnum(NppStatus error) { static const char *_cudaGetErrorEnum(NppStatus error) {
@@ -983,24 +579,12 @@ static const char *_cudaGetErrorEnum(NppStatus error) {
} }
#endif #endif
#ifdef __DRIVER_TYPES_H__
#ifndef DEVICE_RESET
#define DEVICE_RESET cudaDeviceReset();
#endif
#else
#ifndef DEVICE_RESET
#define DEVICE_RESET
#endif
#endif
template <typename T> template <typename T>
void check(T result, char const *const func, const char *const file, void check(T result, char const *const func, const char *const file,
int const line) { int const line) {
if (result) { if (result) {
fprintf(stderr, "CUDA error at %s:%d code=%d(%s) \"%s\" \n", file, line, fprintf(stderr, "CUDA error at %s:%d code=%d(%s) \"%s\" \n", file, line,
static_cast<unsigned int>(result), _cudaGetErrorEnum(result), func); static_cast<unsigned int>(result), _cudaGetErrorEnum(result), func);
DEVICE_RESET
// Make sure we call CUDA Device Reset before exiting
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
} }
@@ -1023,7 +607,6 @@ inline void __getLastCudaError(const char *errorMessage, const char *file,
" %s : (%d) %s.\n", " %s : (%d) %s.\n",
file, line, errorMessage, static_cast<int>(err), file, line, errorMessage, static_cast<int>(err),
cudaGetErrorString(err)); cudaGetErrorString(err));
DEVICE_RESET
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
} }
@@ -1067,18 +650,21 @@ inline int _ConvertSMVer2Cores(int major, int minor) {
} sSMtoCores; } sSMtoCores;
sSMtoCores nGpuArchCoresPerSM[] = { sSMtoCores nGpuArchCoresPerSM[] = {
{0x30, 192}, // Kepler Generation (SM 3.0) GK10x class {0x30, 192},
{0x32, 192}, // Kepler Generation (SM 3.2) GK10x class {0x32, 192},
{0x35, 192}, // Kepler Generation (SM 3.5) GK11x class {0x35, 192},
{0x37, 192}, // Kepler Generation (SM 3.7) GK21x class {0x37, 192},
{0x50, 128}, // Maxwell Generation (SM 5.0) GM10x class {0x50, 128},
{0x52, 128}, // Maxwell Generation (SM 5.2) GM20x class {0x52, 128},
{0x53, 128}, // Maxwell Generation (SM 5.3) GM20x class {0x53, 128},
{0x60, 64}, // Pascal Generation (SM 6.0) GP100 class {0x60, 64},
{0x61, 128}, // Pascal Generation (SM 6.1) GP10x class {0x61, 128},
{0x62, 128}, // Pascal Generation (SM 6.2) GP10x class {0x62, 128},
{0x70, 64}, // Volta Generation (SM 7.0) GV100 class {0x70, 64},
{0x72, 64}, // Volta Generation (SM 7.2) GV11b class {0x72, 64},
{0x75, 64},
{0x80, 64},
{0x86, 128},
{-1, -1}}; {-1, -1}};
int index = 0; int index = 0;
@@ -1099,6 +685,52 @@ inline int _ConvertSMVer2Cores(int major, int minor) {
major, minor, nGpuArchCoresPerSM[index - 1].Cores); major, minor, nGpuArchCoresPerSM[index - 1].Cores);
return nGpuArchCoresPerSM[index - 1].Cores; return nGpuArchCoresPerSM[index - 1].Cores;
} }
inline const char* _ConvertSMVer2ArchName(int major, int minor) {
// Defines for GPU Architecture types (using the SM version to determine
// the GPU Arch name)
typedef struct {
int SM; // 0xMm (hexidecimal notation), M = SM Major version,
// and m = SM minor version
const char* name;
} sSMtoArchName;
sSMtoArchName nGpuArchNameSM[] = {
{0x30, "Kepler"},
{0x32, "Kepler"},
{0x35, "Kepler"},
{0x37, "Kepler"},
{0x50, "Maxwell"},
{0x52, "Maxwell"},
{0x53, "Maxwell"},
{0x60, "Pascal"},
{0x61, "Pascal"},
{0x62, "Pascal"},
{0x70, "Volta"},
{0x72, "Xavier"},
{0x75, "Turing"},
{0x80, "Ampere"},
{0x86, "Ampere"},
{-1, "Graphics Device"}};
int index = 0;
while (nGpuArchNameSM[index].SM != -1) {
if (nGpuArchNameSM[index].SM == ((major << 4) + minor)) {
return nGpuArchNameSM[index].name;
}
index++;
}
// If we don't find the values, we default use the previous one
// to run properly
printf(
"MapSMtoArchName for SM %d.%d is undefined."
" Default to use %s\n",
major, minor, nGpuArchNameSM[index - 1].name);
return nGpuArchNameSM[index - 1].name;
}
// end of GPU Architecture definitions // end of GPU Architecture definitions
#ifdef __CUDA_RUNTIME_H__ #ifdef __CUDA_RUNTIME_H__
@@ -1130,23 +762,24 @@ inline int gpuDeviceInit(int devID) {
return -devID; return -devID;
} }
cudaDeviceProp deviceProp; int computeMode = -1, major = 0, minor = 0;
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, devID)); checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, devID));
checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, devID));
if (deviceProp.computeMode == cudaComputeModeProhibited) { checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, devID));
if (computeMode == cudaComputeModeProhibited) {
fprintf(stderr, fprintf(stderr,
"Error: device is running in <Compute Mode " "Error: device is running in <Compute Mode "
"Prohibited>, no threads can use cudaSetDevice().\n"); "Prohibited>, no threads can use cudaSetDevice().\n");
return -1; return -1;
} }
if (deviceProp.major < 1) { if (major < 1) {
fprintf(stderr, "gpuDeviceInit(): GPU device does not support CUDA.\n"); fprintf(stderr, "gpuDeviceInit(): GPU device does not support CUDA.\n");
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
checkCudaErrors(cudaSetDevice(devID)); checkCudaErrors(cudaSetDevice(devID));
printf("gpuDeviceInit() CUDA Device [%d]: \"%s\n", devID, deviceProp.name); printf("gpuDeviceInit() CUDA Device [%d]: \"%s\n", devID, _ConvertSMVer2ArchName(major, minor));
return devID; return devID;
} }
@@ -1155,11 +788,10 @@ inline int gpuDeviceInit(int devID) {
inline int gpuGetMaxGflopsDeviceId() { inline int gpuGetMaxGflopsDeviceId() {
int current_device = 0, sm_per_multiproc = 0; int current_device = 0, sm_per_multiproc = 0;
int max_perf_device = 0; int max_perf_device = 0;
int device_count = 0, best_SM_arch = 0; int device_count = 0;
int devices_prohibited = 0; int devices_prohibited = 0;
uint64_t max_compute_perf = 0; uint64_t max_compute_perf = 0;
cudaDeviceProp deviceProp;
checkCudaErrors(cudaGetDeviceCount(&device_count)); checkCudaErrors(cudaGetDeviceCount(&device_count));
if (device_count == 0) { if (device_count == 0) {
@@ -1169,21 +801,50 @@ inline int gpuGetMaxGflopsDeviceId() {
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
// Find the best major SM Architecture GPU device // Find the best CUDA capable GPU device
current_device = 0;
while (current_device < device_count) { while (current_device < device_count) {
cudaGetDeviceProperties(&deviceProp, current_device); int computeMode = -1, major = 0, minor = 0;
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, current_device));
checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, current_device));
checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, current_device));
// If this GPU is not running on Compute Mode prohibited, // If this GPU is not running on Compute Mode prohibited,
// then we can add it to the list // then we can add it to the list
if (deviceProp.computeMode != cudaComputeModeProhibited) { if (computeMode != cudaComputeModeProhibited) {
if (deviceProp.major > 0 && deviceProp.major < 9999) { if (major == 9999 && minor == 9999) {
best_SM_arch = MAX(best_SM_arch, deviceProp.major); sm_per_multiproc = 1;
} else {
sm_per_multiproc =
_ConvertSMVer2Cores(major, minor);
}
int multiProcessorCount = 0, clockRate = 0;
checkCudaErrors(cudaDeviceGetAttribute(&multiProcessorCount, cudaDevAttrMultiProcessorCount, current_device));
cudaError_t result = cudaDeviceGetAttribute(&clockRate, cudaDevAttrClockRate, current_device);
if (result != cudaSuccess) {
// If cudaDevAttrClockRate attribute is not supported we
// set clockRate as 1, to consider GPU with most SMs and CUDA Cores.
if(result == cudaErrorInvalidValue) {
clockRate = 1;
}
else {
fprintf(stderr, "CUDA error at %s:%d code=%d(%s) \n", __FILE__, __LINE__,
static_cast<unsigned int>(result), _cudaGetErrorEnum(result));
exit(EXIT_FAILURE);
}
}
uint64_t compute_perf = (uint64_t)multiProcessorCount * sm_per_multiproc * clockRate;
if (compute_perf > max_compute_perf) {
max_compute_perf = compute_perf;
max_perf_device = current_device;
} }
} else { } else {
devices_prohibited++; devices_prohibited++;
} }
current_device++; ++current_device;
} }
if (devices_prohibited == device_count) { if (devices_prohibited == device_count) {
@@ -1193,49 +854,11 @@ inline int gpuGetMaxGflopsDeviceId() {
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
// Find the best CUDA capable GPU device
current_device = 0;
while (current_device < device_count) {
cudaGetDeviceProperties(&deviceProp, current_device);
// If this GPU is not running on Compute Mode prohibited,
// then we can add it to the list
if (deviceProp.computeMode != cudaComputeModeProhibited) {
if (deviceProp.major == 9999 && deviceProp.minor == 9999) {
sm_per_multiproc = 1;
} else {
sm_per_multiproc =
_ConvertSMVer2Cores(deviceProp.major, deviceProp.minor);
}
uint64_t compute_perf = (uint64_t)deviceProp.multiProcessorCount *
sm_per_multiproc * deviceProp.clockRate;
if (compute_perf > max_compute_perf) {
// If we find GPU with SM major > 2, search only these
if (best_SM_arch > 2) {
// If our device==dest_SM_arch, choose this, or else pass
if (deviceProp.major == best_SM_arch) {
max_compute_perf = compute_perf;
max_perf_device = current_device;
}
} else {
max_compute_perf = compute_perf;
max_perf_device = current_device;
}
}
}
++current_device;
}
return max_perf_device; return max_perf_device;
} }
// Initialization code to find the best CUDA Device // Initialization code to find the best CUDA Device
inline int findCudaDevice(int argc, const char **argv) { inline int findCudaDevice(int argc, const char **argv) {
cudaDeviceProp deviceProp;
int devID = 0; int devID = 0;
// If the command-line has a device number specified, use it // If the command-line has a device number specified, use it
@@ -1257,9 +880,12 @@ inline int findCudaDevice(int argc, const char **argv) {
// Otherwise pick the device with highest Gflops/s // Otherwise pick the device with highest Gflops/s
devID = gpuGetMaxGflopsDeviceId(); devID = gpuGetMaxGflopsDeviceId();
checkCudaErrors(cudaSetDevice(devID)); checkCudaErrors(cudaSetDevice(devID));
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, devID)); int major = 0, minor = 0;
printf("GPU Device %d: \"%s\" with compute capability %d.%d\n\n", devID, checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, devID));
deviceProp.name, deviceProp.major, deviceProp.minor); checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, devID));
printf("GPU Device %d: \"%s\" with compute capability %d.%d\n\n",
devID, _ConvertSMVer2ArchName(major, minor), major, minor);
} }
return devID; return devID;
@@ -1270,7 +896,6 @@ inline int findIntegratedGPU() {
int device_count = 0; int device_count = 0;
int devices_prohibited = 0; int devices_prohibited = 0;
cudaDeviceProp deviceProp;
checkCudaErrors(cudaGetDeviceCount(&device_count)); checkCudaErrors(cudaGetDeviceCount(&device_count));
if (device_count == 0) { if (device_count == 0) {
@@ -1280,17 +905,19 @@ inline int findIntegratedGPU() {
// Find the integrated GPU which is compute capable // Find the integrated GPU which is compute capable
while (current_device < device_count) { while (current_device < device_count) {
cudaGetDeviceProperties(&deviceProp, current_device); int computeMode = -1, integrated = -1;
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, current_device));
checkCudaErrors(cudaDeviceGetAttribute(&integrated, cudaDevAttrIntegrated, current_device));
// If GPU is integrated and is not running on Compute Mode prohibited, // If GPU is integrated and is not running on Compute Mode prohibited,
// then cuda can map to GLES resource // then cuda can map to GLES resource
if (deviceProp.integrated && if (integrated && (computeMode != cudaComputeModeProhibited)) {
(deviceProp.computeMode != cudaComputeModeProhibited)) {
checkCudaErrors(cudaSetDevice(current_device)); checkCudaErrors(cudaSetDevice(current_device));
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, current_device));
int major = 0, minor = 0;
checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, current_device));
checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, current_device));
printf("GPU Device %d: \"%s\" with compute capability %d.%d\n\n", printf("GPU Device %d: \"%s\" with compute capability %d.%d\n\n",
current_device, deviceProp.name, deviceProp.major, current_device, _ConvertSMVer2ArchName(major, minor), major, minor);
deviceProp.minor);
return current_device; return current_device;
} else { } else {
@@ -1312,19 +939,18 @@ inline int findIntegratedGPU() {
// General check for CUDA GPU SM Capabilities // General check for CUDA GPU SM Capabilities
inline bool checkCudaCapabilities(int major_version, int minor_version) { inline bool checkCudaCapabilities(int major_version, int minor_version) {
cudaDeviceProp deviceProp;
deviceProp.major = 0;
deviceProp.minor = 0;
int dev; int dev;
int major = 0, minor = 0;
checkCudaErrors(cudaGetDevice(&dev)); checkCudaErrors(cudaGetDevice(&dev));
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, dev)); checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, dev));
checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, dev));
if ((deviceProp.major > major_version) || if ((major > major_version) ||
(deviceProp.major == major_version && (major == major_version &&
deviceProp.minor >= minor_version)) { minor >= minor_version)) {
printf(" Device %d: <%16s >, Compute SM %d.%d detected\n", dev, printf(" Device %d: <%16s >, Compute SM %d.%d detected\n", dev,
deviceProp.name, deviceProp.major, deviceProp.minor); _ConvertSMVer2ArchName(major, minor), major, minor);
return true; return true;
} else { } else {
printf( printf(

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -33,8 +33,10 @@
#include <stdio.h> #include <stdio.h>
#include <stdlib.h> #include <stdlib.h>
#include <string.h> #include <string.h>
#include <iostream>
#include <cstring>
#include <sstream>
#include <drvapi_error_string.h>
#include <helper_string.h> #include <helper_string.h>
#ifndef MAX #ifndef MAX
@@ -66,48 +68,22 @@ inline int ftoi(float value) {
// These are the inline versions for all of the SDK helper functions // These are the inline versions for all of the SDK helper functions
inline void __checkCudaErrors(CUresult err, const char *file, const int line) { inline void __checkCudaErrors(CUresult err, const char *file, const int line) {
if (CUDA_SUCCESS != err) { if (CUDA_SUCCESS != err) {
const char *errorStr = NULL;
cuGetErrorString(err, &errorStr);
fprintf(stderr, fprintf(stderr,
"checkCudaErrors() Driver API error = %04d \"%s\" from file <%s>, " "checkCudaErrors() Driver API error = %04d \"%s\" from file <%s>, "
"line %i.\n", "line %i.\n",
err, getCudaDrvErrorString(err), file, line); err, errorStr, file, line);
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
} }
#endif #endif
#ifdef getLastCudaDrvErrorMsg
#undef getLastCudaDrvErrorMsg
#endif
#define getLastCudaDrvErrorMsg(msg) \
__getLastCudaDrvErrorMsg(msg, __FILE__, __LINE__)
inline void __getLastCudaDrvErrorMsg(const char *msg, const char *file,
const int line) {
CUresult err = cuCtxSynchronize();
if (CUDA_SUCCESS != err) {
fprintf(stderr, "getLastCudaDrvErrorMsg -> %s", msg);
fprintf(stderr,
"getLastCudaDrvErrorMsg -> cuCtxSynchronize API error = %04d "
"\"%s\" in file <%s>, line %i.\n",
err, getCudaDrvErrorString(err), file, line);
exit(EXIT_FAILURE);
}
}
// This function wraps the CUDA Driver API into a template function // This function wraps the CUDA Driver API into a template function
template <class T> template <class T>
inline void getCudaAttribute(T *attribute, CUdevice_attribute device_attribute, inline void getCudaAttribute(T *attribute, CUdevice_attribute device_attribute,
int device) { int device) {
CUresult error_result = checkCudaErrors(cuDeviceGetAttribute(attribute, device_attribute, device));
cuDeviceGetAttribute(attribute, device_attribute, device);
if (error_result != CUDA_SUCCESS) {
printf("cuDeviceGetAttribute returned %d\n-> %s\n",
static_cast<int>(error_result), getCudaDrvErrorString(error_result));
exit(EXIT_SUCCESS);
}
} }
#endif #endif
@@ -122,18 +98,21 @@ inline int _ConvertSMVer2CoresDRV(int major, int minor) {
} sSMtoCores; } sSMtoCores;
sSMtoCores nGpuArchCoresPerSM[] = { sSMtoCores nGpuArchCoresPerSM[] = {
{0x30, 192}, // Kepler Generation (SM 3.0) GK10x class {0x30, 192},
{0x32, 192}, // Kepler Generation (SM 3.2) GK10x class {0x32, 192},
{0x35, 192}, // Kepler Generation (SM 3.5) GK11x class {0x35, 192},
{0x37, 192}, // Kepler Generation (SM 3.7) GK21x class {0x37, 192},
{0x50, 128}, // Maxwell Generation (SM 5.0) GM10x class {0x50, 128},
{0x52, 128}, // Maxwell Generation (SM 5.2) GM20x class {0x52, 128},
{0x53, 128}, // Maxwell Generation (SM 5.3) GM20x class {0x53, 128},
{0x60, 64}, // Pascal Generation (SM 6.0) GP100 class {0x60, 64},
{0x61, 128}, // Pascal Generation (SM 6.1) GP10x class {0x61, 128},
{0x62, 128}, // Pascal Generation (SM 6.2) GP10x class {0x62, 128},
{0x70, 64}, // Volta Generation (SM 7.0) GV100 class {0x70, 64},
{0x72, 64}, // Volta Generation (SM 7.2) GV11b class {0x72, 64},
{0x75, 64},
{0x80, 64},
{0x86, 128},
{-1, -1}}; {-1, -1}};
int index = 0; int index = 0;
@@ -160,11 +139,9 @@ inline int _ConvertSMVer2CoresDRV(int major, int minor) {
inline int gpuDeviceInitDRV(int ARGC, const char **ARGV) { inline int gpuDeviceInitDRV(int ARGC, const char **ARGV) {
int cuDevice = 0; int cuDevice = 0;
int deviceCount = 0; int deviceCount = 0;
CUresult err = cuInit(0); checkCudaErrors(cuInit(0));
if (CUDA_SUCCESS == err) {
checkCudaErrors(cuDeviceGetCount(&deviceCount)); checkCudaErrors(cuDeviceGetCount(&deviceCount));
}
if (deviceCount == 0) { if (deviceCount == 0) {
fprintf(stderr, "cudaDeviceInit error: no devices supporting CUDA\n"); fprintf(stderr, "cudaDeviceInit error: no devices supporting CUDA\n");
@@ -191,7 +168,7 @@ inline int gpuDeviceInitDRV(int ARGC, const char **ARGV) {
checkCudaErrors(cuDeviceGet(&cuDevice, dev)); checkCudaErrors(cuDeviceGet(&cuDevice, dev));
char name[100]; char name[100];
cuDeviceGetName(name, 100, cuDevice); checkCudaErrors(cuDeviceGetName(name, 100, cuDevice));
int computeMode; int computeMode;
getCudaAttribute<int>(&computeMode, CU_DEVICE_ATTRIBUTE_COMPUTE_MODE, dev); getCudaAttribute<int>(&computeMode, CU_DEVICE_ATTRIBUTE_COMPUTE_MODE, dev);
@@ -217,7 +194,6 @@ inline int gpuGetMaxGflopsDeviceIdDRV() {
int device_count = 0; int device_count = 0;
int sm_per_multiproc = 0; int sm_per_multiproc = 0;
unsigned long long max_compute_perf = 0; unsigned long long max_compute_perf = 0;
int best_SM_arch = 0;
int major = 0; int major = 0;
int minor = 0; int minor = 0;
int multiProcessorCount; int multiProcessorCount;
@@ -233,19 +209,6 @@ inline int gpuGetMaxGflopsDeviceIdDRV() {
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
// Find the best major SM Architecture GPU device
while (current_device < device_count) {
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, current_device));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, current_device));
if (major > 0 && major < 9999) {
best_SM_arch = MAX(best_SM_arch, major);
}
current_device++;
}
// Find the best CUDA capable GPU device // Find the best CUDA capable GPU device
current_device = 0; current_device = 0;
@@ -276,18 +239,9 @@ inline int gpuGetMaxGflopsDeviceIdDRV() {
clockRate); clockRate);
if (compute_perf > max_compute_perf) { if (compute_perf > max_compute_perf) {
// If we find GPU with SM major > 2, search only these
if (best_SM_arch > 2) {
// If our device==dest_SM_arch, choose this, or else pass
if (major == best_SM_arch) {
max_compute_perf = compute_perf; max_compute_perf = compute_perf;
max_perf_device = current_device; max_perf_device = current_device;
} }
} else {
max_compute_perf = compute_perf;
max_perf_device = current_device;
}
}
} else { } else {
devices_prohibited++; devices_prohibited++;
} }
@@ -413,7 +367,39 @@ inline bool checkCudaCapabilitiesDRV(int major_version, int minor_version,
} }
} }
#endif #endif
bool inline findFatbinPath(const char *module_file, std::string &module_path, char **argv, std::ostringstream &ostrm)
{
char *actual_path = sdkFindFilePath(module_file, argv[0]);
// end of CUDA Helper Functions if (actual_path)
{
module_path = actual_path;
}
else
{
printf("> findModulePath file not found: <%s> \n", module_file);
return false;
}
if (module_path.empty())
{
printf("> findModulePath could not find file: <%s> \n", module_file);
return false;
}
else
{
printf("> findModulePath found file at <%s>\n", module_path.c_str());
if (module_path.rfind("fatbin") != std::string::npos)
{
std::ifstream fileIn(module_path.c_str(), std::ios::binary);
ostrm << fileIn.rdbuf();
fileIn.close();
}
return true;
}
}
// end of CUDA Helper Functions
#endif // COMMON_HELPER_CUDA_DRVAPI_H_ #endif // COMMON_HELPER_CUDA_DRVAPI_H_

166
Common/helper_cusolver.h Normal file
View File

@@ -0,0 +1,166 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef HELPER_CUSOLVER
#define HELPER_CUSOLVER
#include <ctype.h>
#include <cuda_runtime.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "cusparse.h"
#define SWITCH_CHAR '-'
struct testOpts {
char *sparse_mat_filename; // by switch -F<filename>
const char *testFunc; // by switch -R<name>
const char *reorder; // by switch -P<name>
int lda; // by switch -lda<int>
};
double vec_norminf(int n, const double *x) {
double norminf = 0;
for (int j = 0; j < n; j++) {
double x_abs = fabs(x[j]);
norminf = (norminf > x_abs) ? norminf : x_abs;
}
return norminf;
}
/*
* |A| = max { |A|*ones(m,1) }
*/
double mat_norminf(int m, int n, const double *A, int lda) {
double norminf = 0;
for (int i = 0; i < m; i++) {
double sum = 0.0;
for (int j = 0; j < n; j++) {
double A_abs = fabs(A[i + j * lda]);
sum += A_abs;
}
norminf = (norminf > sum) ? norminf : sum;
}
return norminf;
}
/*
* |A| = max { |A|*ones(m,1) }
*/
double csr_mat_norminf(int m, int n, int nnzA, const cusparseMatDescr_t descrA,
const double *csrValA, const int *csrRowPtrA,
const int *csrColIndA) {
const int baseA =
(CUSPARSE_INDEX_BASE_ONE == cusparseGetMatIndexBase(descrA)) ? 1 : 0;
double norminf = 0;
for (int i = 0; i < m; i++) {
double sum = 0.0;
const int start = csrRowPtrA[i] - baseA;
const int end = csrRowPtrA[i + 1] - baseA;
for (int colidx = start; colidx < end; colidx++) {
// const int j = csrColIndA[colidx] - baseA;
double A_abs = fabs(csrValA[colidx]);
sum += A_abs;
}
norminf = (norminf > sum) ? norminf : sum;
}
return norminf;
}
void display_matrix(int m, int n, int nnzA, const cusparseMatDescr_t descrA,
const double *csrValA, const int *csrRowPtrA,
const int *csrColIndA) {
const int baseA =
(CUSPARSE_INDEX_BASE_ONE == cusparseGetMatIndexBase(descrA)) ? 1 : 0;
printf("m = %d, n = %d, nnz = %d, matlab base-1\n", m, n, nnzA);
for (int row = 0; row < m; row++) {
const int start = csrRowPtrA[row] - baseA;
const int end = csrRowPtrA[row + 1] - baseA;
for (int colidx = start; colidx < end; colidx++) {
const int col = csrColIndA[colidx] - baseA;
double Areg = csrValA[colidx];
printf("A(%d, %d) = %20.16E\n", row + 1, col + 1, Areg);
}
}
}
#if defined(_WIN32)
#if !defined(WIN32_LEAN_AND_MEAN)
#define WIN32_LEAN_AND_MEAN
#endif
#include <windows.h>
double second(void) {
LARGE_INTEGER t;
static double oofreq;
static int checkedForHighResTimer;
static BOOL hasHighResTimer;
if (!checkedForHighResTimer) {
hasHighResTimer = QueryPerformanceFrequency(&t);
oofreq = 1.0 / (double)t.QuadPart;
checkedForHighResTimer = 1;
}
if (hasHighResTimer) {
QueryPerformanceCounter(&t);
return (double)t.QuadPart * oofreq;
} else {
return (double)GetTickCount() / 1000.0;
}
}
#elif defined(__linux__) || defined(__QNX__)
#include <stddef.h>
#include <sys/resource.h>
#include <sys/time.h>
double second(void) {
struct timeval tv;
gettimeofday(&tv, NULL);
return (double)tv.tv_sec + (double)tv.tv_usec / 1000000.0;
}
#elif defined(__APPLE__)
#include <stddef.h>
#include <sys/resource.h>
#include <sys/sysctl.h>
#include <sys/time.h>
#include <sys/types.h>
double second(void) {
struct timeval tv;
gettimeofday(&tv, NULL);
return (double)tv.tv_sec + (double)tv.tv_usec / 1000000.0;
}
#else
#error unsupported platform
#endif
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

267
Common/helper_gl.h Normal file
View File

@@ -0,0 +1,267 @@
/* Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
// These are helper functions for the SDK samples (OpenGL)
#ifndef HELPER_GL_H
#define HELPER_GL_H
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
#include <GL/glew.h>
#endif
#if defined(__APPLE__) || defined(MACOSX)
#include <OpenGL/gl.h>
#else
#include <GL/gl.h>
#ifdef __linux__
#include <GL/glx.h>
#endif /* __linux__ */
#endif
#include <iostream>
#include <cstdio>
#include <string>
#include <sstream>
#include <algorithm>
#include <iterator>
#include <vector>
#include <assert.h>
/* Prototypes */
namespace __HelperGL {
static int isGLVersionSupported(unsigned reqMajor, unsigned reqMinor);
static int areGLExtensionsSupported(const std::string &);
#ifdef __linux__
#ifndef HELPERGL_EXTERN_GL_FUNC_IMPLEMENTATION
#define USE_GL_FUNC(name, proto) proto name = (proto) glXGetProcAddress ((const GLubyte *)#name)
#else
#define USE_GL_FUNC(name, proto) extern proto name
#endif
USE_GL_FUNC(glBindBuffer, PFNGLBINDBUFFERPROC);
USE_GL_FUNC(glDeleteBuffers, PFNGLDELETEBUFFERSPROC);
USE_GL_FUNC(glBufferData, PFNGLBUFFERDATAPROC);
USE_GL_FUNC(glBufferSubData, PFNGLBUFFERSUBDATAPROC);
USE_GL_FUNC(glGenBuffers, PFNGLGENBUFFERSPROC);
USE_GL_FUNC(glCreateProgram, PFNGLCREATEPROGRAMPROC);
USE_GL_FUNC(glBindProgramARB, PFNGLBINDPROGRAMARBPROC);
USE_GL_FUNC(glGenProgramsARB, PFNGLGENPROGRAMSARBPROC);
USE_GL_FUNC(glDeleteProgramsARB, PFNGLDELETEPROGRAMSARBPROC);
USE_GL_FUNC(glDeleteProgram, PFNGLDELETEPROGRAMPROC);
USE_GL_FUNC(glGetProgramInfoLog, PFNGLGETPROGRAMINFOLOGPROC);
USE_GL_FUNC(glGetProgramiv, PFNGLGETPROGRAMIVPROC);
USE_GL_FUNC(glProgramParameteriEXT, PFNGLPROGRAMPARAMETERIEXTPROC);
USE_GL_FUNC(glProgramStringARB, PFNGLPROGRAMSTRINGARBPROC);
USE_GL_FUNC(glUnmapBuffer, PFNGLUNMAPBUFFERPROC);
USE_GL_FUNC(glMapBuffer, PFNGLMAPBUFFERPROC);
USE_GL_FUNC(glGetBufferParameteriv, PFNGLGETBUFFERPARAMETERIVPROC);
USE_GL_FUNC(glLinkProgram, PFNGLLINKPROGRAMPROC);
USE_GL_FUNC(glUseProgram, PFNGLUSEPROGRAMPROC);
USE_GL_FUNC(glAttachShader, PFNGLATTACHSHADERPROC);
USE_GL_FUNC(glCreateShader, PFNGLCREATESHADERPROC);
USE_GL_FUNC(glShaderSource, PFNGLSHADERSOURCEPROC);
USE_GL_FUNC(glCompileShader, PFNGLCOMPILESHADERPROC);
USE_GL_FUNC(glDeleteShader, PFNGLDELETESHADERPROC);
USE_GL_FUNC(glGetShaderInfoLog, PFNGLGETSHADERINFOLOGPROC);
USE_GL_FUNC(glGetShaderiv, PFNGLGETSHADERIVPROC);
USE_GL_FUNC(glUniform1i, PFNGLUNIFORM1IPROC);
USE_GL_FUNC(glUniform1f, PFNGLUNIFORM1FPROC);
USE_GL_FUNC(glUniform2f, PFNGLUNIFORM2FPROC);
USE_GL_FUNC(glUniform3f, PFNGLUNIFORM3FPROC);
USE_GL_FUNC(glUniform4f, PFNGLUNIFORM4FPROC);
USE_GL_FUNC(glUniform1fv, PFNGLUNIFORM1FVPROC);
USE_GL_FUNC(glUniform2fv, PFNGLUNIFORM2FVPROC);
USE_GL_FUNC(glUniform3fv, PFNGLUNIFORM3FVPROC);
USE_GL_FUNC(glUniform4fv, PFNGLUNIFORM4FVPROC);
USE_GL_FUNC(glUniformMatrix4fv, PFNGLUNIFORMMATRIX4FVPROC);
USE_GL_FUNC(glSecondaryColor3fv, PFNGLSECONDARYCOLOR3FVPROC);
USE_GL_FUNC(glGetUniformLocation, PFNGLGETUNIFORMLOCATIONPROC);
USE_GL_FUNC(glGenFramebuffersEXT, PFNGLGENFRAMEBUFFERSEXTPROC);
USE_GL_FUNC(glBindFramebufferEXT, PFNGLBINDFRAMEBUFFEREXTPROC);
USE_GL_FUNC(glDeleteFramebuffersEXT, PFNGLDELETEFRAMEBUFFERSEXTPROC);
USE_GL_FUNC(glCheckFramebufferStatusEXT, PFNGLCHECKFRAMEBUFFERSTATUSEXTPROC);
USE_GL_FUNC(glGetFramebufferAttachmentParameterivEXT, PFNGLGETFRAMEBUFFERATTACHMENTPARAMETERIVEXTPROC);
USE_GL_FUNC(glFramebufferTexture1DEXT, PFNGLFRAMEBUFFERTEXTURE1DEXTPROC);
USE_GL_FUNC(glFramebufferTexture2DEXT, PFNGLFRAMEBUFFERTEXTURE2DEXTPROC);
USE_GL_FUNC(glFramebufferTexture3DEXT, PFNGLFRAMEBUFFERTEXTURE3DEXTPROC);
USE_GL_FUNC(glGenerateMipmapEXT, PFNGLGENERATEMIPMAPEXTPROC);
USE_GL_FUNC(glGenRenderbuffersEXT, PFNGLGENRENDERBUFFERSEXTPROC);
USE_GL_FUNC(glDeleteRenderbuffersEXT, PFNGLDELETERENDERBUFFERSEXTPROC);
USE_GL_FUNC(glBindRenderbufferEXT, PFNGLBINDRENDERBUFFEREXTPROC);
USE_GL_FUNC(glRenderbufferStorageEXT, PFNGLRENDERBUFFERSTORAGEEXTPROC);
USE_GL_FUNC(glFramebufferRenderbufferEXT, PFNGLFRAMEBUFFERRENDERBUFFEREXTPROC);
USE_GL_FUNC(glClampColorARB, PFNGLCLAMPCOLORARBPROC);
USE_GL_FUNC(glBindFragDataLocationEXT, PFNGLBINDFRAGDATALOCATIONEXTPROC);
#if !defined(GLX_EXTENSION_NAME) || !defined(GL_VERSION_1_3)
USE_GL_FUNC(glActiveTexture, PFNGLACTIVETEXTUREPROC);
USE_GL_FUNC(glClientActiveTexture, PFNGLACTIVETEXTUREPROC);
#endif
#undef USE_GL_FUNC
#endif /*__linux__ */
}
namespace __HelperGL {
namespace __Int {
static std::vector<std::string> split(const std::string &str)
{
std::istringstream ss(str);
std::istream_iterator<std::string> it(ss);
return std::vector<std::string> (it, std::istream_iterator<std::string>());
}
/* Sort the vector passed by reference */
template<typename T> static inline void sort(std::vector<T> &a)
{
std::sort(a.begin(), a.end());
}
/* Compare two vectors */
template<typename T> static int equals(std::vector<T> a, std::vector<T> b)
{
if (a.size() != b.size()) return 0;
sort(a);
sort(b);
return std::equal(a.begin(), a.end(), b.begin());
}
template<typename T> static std::vector<T> getIntersection(std::vector<T> a, std::vector<T> b)
{
sort(a);
sort(b);
std::vector<T> rc;
std::set_intersection(a.begin(), a.end(), b.begin(), b.end(),
std::back_inserter<std::vector<std::string> >(rc));
return rc;
}
static std::vector<std::string> getGLExtensions()
{
std::string extensionsStr( (const char *)glGetString(GL_EXTENSIONS));
return split (extensionsStr);
}
}
static int areGLExtensionsSupported(const std::string &extensions)
{
std::vector<std::string> all = __Int::getGLExtensions();
std::vector<std::string> requested = __Int::split(extensions);
std::vector<std::string> matched = __Int::getIntersection(all, requested);
return __Int::equals(matched, requested);
}
static int isGLVersionSupported(unsigned reqMajor, unsigned reqMinor)
{
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
if (glewInit() != GLEW_OK)
{
std::cerr << "glewInit() failed!" << std::endl;
return 0;
}
#endif
std::string version ((const char *) glGetString (GL_VERSION));
std::stringstream stream (version);
unsigned major, minor;
char dot;
stream >> major >> dot >> minor;
assert (dot == '.');
return major > reqMajor || (major == reqMajor && minor >= reqMinor);
}
static inline const char* glErrorToString(GLenum err)
{
#define CASE_RETURN_MACRO(arg) case arg: return #arg
switch(err)
{
CASE_RETURN_MACRO(GL_NO_ERROR);
CASE_RETURN_MACRO(GL_INVALID_ENUM);
CASE_RETURN_MACRO(GL_INVALID_VALUE);
CASE_RETURN_MACRO(GL_INVALID_OPERATION);
CASE_RETURN_MACRO(GL_OUT_OF_MEMORY);
CASE_RETURN_MACRO(GL_STACK_UNDERFLOW);
CASE_RETURN_MACRO(GL_STACK_OVERFLOW);
#ifdef GL_INVALID_FRAMEBUFFER_OPERATION
CASE_RETURN_MACRO(GL_INVALID_FRAMEBUFFER_OPERATION);
#endif
default: break;
}
#undef CASE_RETURN_MACRO
return "*UNKNOWN*";
}
////////////////////////////////////////////////////////////////////////////
//! Check for OpenGL error
//! @return bool if no GL error has been encountered, otherwise 0
//! @param file __FILE__ macro
//! @param line __LINE__ macro
//! @note The GL error is listed on stderr
//! @note This function should be used via the CHECK_ERROR_GL() macro
////////////////////////////////////////////////////////////////////////////
inline bool sdkCheckErrorGL(const char *file, const int line)
{
bool ret_val = true;
// check for error
GLenum gl_error = glGetError();
if (gl_error != GL_NO_ERROR)
{
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
char tmpStr[512];
// NOTE: "%s(%i) : " allows Visual Studio to directly jump to the file at the right line
// when the user double clicks on the error line in the Output pane. Like any compile error.
sprintf_s(tmpStr, 255, "\n%s(%i) : GL Error : %s\n\n", file, line, glErrorToString(gl_error));
fprintf(stderr, "%s", tmpStr);
#endif
fprintf(stderr, "GL Error in file '%s' in line %d :\n", file, line);
fprintf(stderr, "%s\n", glErrorToString(gl_error));
ret_val = false;
}
return ret_val;
}
#define SDK_CHECK_ERROR_GL() \
if( false == sdkCheckErrorGL( __FILE__, __LINE__)) { \
exit(EXIT_FAILURE); \
}
} /* of namespace __HelperGL*/
using namespace __HelperGL;
#endif /*HELPER_GL_H*/

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

1469
Common/helper_math.h Normal file

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,543 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "helper_multiprocess.h"
#include <cstdlib>
#include <string>
int sharedMemoryCreate(const char *name, size_t sz, sharedMemoryInfo *info) {
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
info->size = sz;
info->shmHandle = CreateFileMapping(INVALID_HANDLE_VALUE, NULL,
PAGE_READWRITE, 0, (DWORD)sz, name);
if (info->shmHandle == 0) {
return GetLastError();
}
info->addr = MapViewOfFile(info->shmHandle, FILE_MAP_ALL_ACCESS, 0, 0, sz);
if (info->addr == NULL) {
return GetLastError();
}
return 0;
#else
int status = 0;
info->size = sz;
info->shmFd = shm_open(name, O_RDWR | O_CREAT, 0777);
if (info->shmFd < 0) {
return errno;
}
status = ftruncate(info->shmFd, sz);
if (status != 0) {
return status;
}
info->addr = mmap(0, sz, PROT_READ | PROT_WRITE, MAP_SHARED, info->shmFd, 0);
if (info->addr == NULL) {
return errno;
}
return 0;
#endif
}
int sharedMemoryOpen(const char *name, size_t sz, sharedMemoryInfo *info) {
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
info->size = sz;
info->shmHandle = OpenFileMapping(FILE_MAP_ALL_ACCESS, FALSE, name);
if (info->shmHandle == 0) {
return GetLastError();
}
info->addr = MapViewOfFile(info->shmHandle, FILE_MAP_ALL_ACCESS, 0, 0, sz);
if (info->addr == NULL) {
return GetLastError();
}
return 0;
#else
info->size = sz;
info->shmFd = shm_open(name, O_RDWR, 0777);
if (info->shmFd < 0) {
return errno;
}
info->addr = mmap(0, sz, PROT_READ | PROT_WRITE, MAP_SHARED, info->shmFd, 0);
if (info->addr == NULL) {
return errno;
}
return 0;
#endif
}
void sharedMemoryClose(sharedMemoryInfo *info) {
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
if (info->addr) {
UnmapViewOfFile(info->addr);
}
if (info->shmHandle) {
CloseHandle(info->shmHandle);
}
#else
if (info->addr) {
munmap(info->addr, info->size);
}
if (info->shmFd) {
close(info->shmFd);
}
#endif
}
int spawnProcess(Process *process, const char *app, char *const *args) {
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
STARTUPINFO si = {0};
BOOL status;
size_t arglen = 0;
size_t argIdx = 0;
std::string arg_string;
memset(process, 0, sizeof(*process));
while (*args) {
arg_string.append(*args).append(1, ' ');
args++;
}
status = CreateProcess(app, LPSTR(arg_string.c_str()), NULL, NULL, FALSE, 0,
NULL, NULL, &si, process);
return status ? 0 : GetLastError();
#else
*process = fork();
if (*process == 0) {
if (0 > execvp(app, args)) {
return errno;
}
} else if (*process < 0) {
return errno;
}
return 0;
#endif
}
int waitProcess(Process *process) {
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
DWORD exitCode;
WaitForSingleObject(process->hProcess, INFINITE);
GetExitCodeProcess(process->hProcess, &exitCode);
CloseHandle(process->hProcess);
CloseHandle(process->hThread);
return (int)exitCode;
#else
int status = 0;
do {
if (0 > waitpid(*process, &status, 0)) {
return errno;
}
} while (!WIFEXITED(status));
return WEXITSTATUS(status);
#endif
}
#if defined(__linux__)
int ipcCreateSocket(ipcHandle *&handle, const char *name,
const std::vector<Process> &processes) {
int server_fd;
struct sockaddr_un servaddr;
handle = new ipcHandle;
memset(handle, 0, sizeof(*handle));
handle->socket = -1;
handle->socketName = NULL;
// Creating socket file descriptor
if ((server_fd = socket(AF_UNIX, SOCK_DGRAM, 0)) == 0) {
perror("IPC failure: Socket creation failed");
return -1;
}
unlink(name);
bzero(&servaddr, sizeof(servaddr));
servaddr.sun_family = AF_UNIX;
size_t len = strlen(name);
if (len > (sizeof(servaddr.sun_path) - 1)) {
perror("IPC failure: Cannot bind provided name to socket. Name too large");
return -1;
}
strncpy(servaddr.sun_path, name, len);
if (bind(server_fd, (struct sockaddr *)&servaddr, SUN_LEN(&servaddr)) < 0) {
perror("IPC failure: Binding socket failed");
return -1;
}
handle->socketName = new char[strlen(name) + 1];
strcpy(handle->socketName, name);
handle->socket = server_fd;
return 0;
}
int ipcOpenSocket(ipcHandle *&handle) {
int sock = 0;
struct sockaddr_un cliaddr;
handle = new ipcHandle;
memset(handle, 0, sizeof(*handle));
if ((sock = socket(AF_UNIX, SOCK_DGRAM, 0)) < 0) {
perror("IPC failure:Socket creation error");
return -1;
}
bzero(&cliaddr, sizeof(cliaddr));
cliaddr.sun_family = AF_UNIX;
char temp[10];
// Create unique name for the socket.
sprintf(temp, "%u", getpid());
strcpy(cliaddr.sun_path, temp);
if (bind(sock, (struct sockaddr *)&cliaddr, sizeof(cliaddr)) < 0) {
perror("IPC failure: Binding socket failed");
return -1;
}
handle->socket = sock;
handle->socketName = new char[strlen(temp) + 1];
strcpy(handle->socketName, temp);
return 0;
}
int ipcCloseSocket(ipcHandle *handle) {
if (!handle) {
return -1;
}
if (handle->socketName) {
unlink(handle->socketName);
delete[] handle->socketName;
}
close(handle->socket);
delete handle;
return 0;
}
int ipcRecvShareableHandle(ipcHandle *handle, ShareableHandle *shHandle) {
struct msghdr msg = {0};
struct iovec iov[1];
struct cmsghdr cm;
// Union to guarantee alignment requirements for control array
union {
struct cmsghdr cm;
char control[CMSG_SPACE(sizeof(int))];
} control_un;
struct cmsghdr *cmptr;
ssize_t n;
int receivedfd;
char dummy_buffer[1];
ssize_t sendResult;
msg.msg_control = control_un.control;
msg.msg_controllen = sizeof(control_un.control);
iov[0].iov_base = (void *)dummy_buffer;
iov[0].iov_len = sizeof(dummy_buffer);
msg.msg_iov = iov;
msg.msg_iovlen = 1;
if ((n = recvmsg(handle->socket, &msg, 0)) <= 0) {
perror("IPC failure: Receiving data over socket failed");
return -1;
}
if (((cmptr = CMSG_FIRSTHDR(&msg)) != NULL) &&
(cmptr->cmsg_len == CMSG_LEN(sizeof(int)))) {
if ((cmptr->cmsg_level != SOL_SOCKET) || (cmptr->cmsg_type != SCM_RIGHTS)) {
return -1;
}
memmove(&receivedfd, CMSG_DATA(cmptr), sizeof(receivedfd));
*(int *)shHandle = receivedfd;
} else {
return -1;
}
return 0;
}
int ipcRecvDataFromClient(ipcHandle *serverHandle, void *data, size_t size) {
ssize_t readResult;
struct sockaddr_un cliaddr;
socklen_t len = sizeof(cliaddr);
readResult = recvfrom(serverHandle->socket, data, size, 0,
(struct sockaddr *)&cliaddr, &len);
if (readResult == -1) {
perror("IPC failure: Receiving data over socket failed");
return -1;
}
return 0;
}
int ipcSendDataToServer(ipcHandle *handle, const char *serverName,
const void *data, size_t size) {
ssize_t sendResult;
struct sockaddr_un serveraddr;
bzero(&serveraddr, sizeof(serveraddr));
serveraddr.sun_family = AF_UNIX;
strncpy(serveraddr.sun_path, serverName, sizeof(serveraddr.sun_path) - 1);
sendResult = sendto(handle->socket, data, size, 0,
(struct sockaddr *)&serveraddr, sizeof(serveraddr));
if (sendResult <= 0) {
perror("IPC failure: Sending data over socket failed");
}
return 0;
}
int ipcSendShareableHandle(ipcHandle *handle,
const std::vector<ShareableHandle> &shareableHandles,
Process process, int data) {
struct msghdr msg;
struct iovec iov[1];
union {
struct cmsghdr cm;
char control[CMSG_SPACE(sizeof(int))];
} control_un;
struct cmsghdr *cmptr;
ssize_t readResult;
struct sockaddr_un cliaddr;
socklen_t len = sizeof(cliaddr);
// Construct client address to send this SHareable handle to
bzero(&cliaddr, sizeof(cliaddr));
cliaddr.sun_family = AF_UNIX;
char temp[10];
sprintf(temp, "%u", process);
strcpy(cliaddr.sun_path, temp);
len = sizeof(cliaddr);
// Send corresponding shareable handle to the client
int sendfd = (int)shareableHandles[data];
msg.msg_control = control_un.control;
msg.msg_controllen = sizeof(control_un.control);
cmptr = CMSG_FIRSTHDR(&msg);
cmptr->cmsg_len = CMSG_LEN(sizeof(int));
cmptr->cmsg_level = SOL_SOCKET;
cmptr->cmsg_type = SCM_RIGHTS;
memmove(CMSG_DATA(cmptr), &sendfd, sizeof(sendfd));
msg.msg_name = (void *)&cliaddr;
msg.msg_namelen = sizeof(struct sockaddr_un);
iov[0].iov_base = (void *)"";
iov[0].iov_len = 1;
msg.msg_iov = iov;
msg.msg_iovlen = 1;
ssize_t sendResult = sendmsg(handle->socket, &msg, 0);
if (sendResult <= 0) {
perror("IPC failure: Sending data over socket failed");
return -1;
}
return 0;
}
int ipcSendShareableHandles(
ipcHandle *handle, const std::vector<ShareableHandle> &shareableHandles,
const std::vector<Process> &processes) {
// Send all shareable handles to every single process.
for (int i = 0; i < shareableHandles.size(); i++) {
for (int j = 0; j < processes.size(); j++) {
checkIpcErrors(
ipcSendShareableHandle(handle, shareableHandles, processes[j], i));
}
}
return 0;
}
int ipcRecvShareableHandles(ipcHandle *handle,
std::vector<ShareableHandle> &shareableHandles) {
for (int i = 0; i < shareableHandles.size(); i++) {
checkIpcErrors(ipcRecvShareableHandle(handle, &shareableHandles[i]));
}
return 0;
}
int ipcCloseShareableHandle(ShareableHandle shHandle) {
return close(shHandle);
}
#elif defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
// Generic name to build individual Mailslot names by appending process ids.
LPTSTR SlotName = (LPTSTR)TEXT("\\\\.\\mailslot\\sample_mailslot_");
int ipcCreateSocket(ipcHandle *&handle, const char *name,
const std::vector<Process> &processes) {
handle = new ipcHandle;
handle->hMailslot.resize(processes.size());
// Open Mailslots of all clients and store respective handles.
for (int i = 0; i < handle->hMailslot.size(); ++i) {
std::basic_string<TCHAR> childSlotName(SlotName);
char tempBuf[20];
_itoa_s(processes[i].dwProcessId, tempBuf, 10);
childSlotName += TEXT(tempBuf);
HANDLE hFile =
CreateFile(TEXT(childSlotName.c_str()), GENERIC_WRITE, FILE_SHARE_READ,
(LPSECURITY_ATTRIBUTES)NULL, OPEN_EXISTING,
FILE_ATTRIBUTE_NORMAL, (HANDLE)NULL);
if (hFile == INVALID_HANDLE_VALUE) {
printf("IPC failure: Opening Mailslot by CreateFile failed with %d\n",
GetLastError());
return -1;
}
handle->hMailslot[i] = hFile;
}
return 0;
}
int ipcOpenSocket(ipcHandle *&handle) {
handle = new ipcHandle;
HANDLE hSlot;
std::basic_string<TCHAR> clientSlotName(SlotName);
char tempBuf[20];
_itoa_s(GetCurrentProcessId(), tempBuf, 10);
clientSlotName += TEXT(tempBuf);
hSlot = CreateMailslot((LPSTR)clientSlotName.c_str(), 0,
MAILSLOT_WAIT_FOREVER, (LPSECURITY_ATTRIBUTES)NULL);
if (hSlot == INVALID_HANDLE_VALUE) {
printf("IPC failure: CreateMailslot failed for client with %d\n",
GetLastError());
return -1;
}
handle->hMailslot.push_back(hSlot);
return 0;
}
int ipcSendData(HANDLE mailslot, const void *data, size_t sz) {
BOOL result;
DWORD cbWritten;
result = WriteFile(mailslot, data, (DWORD)sz, &cbWritten, (LPOVERLAPPED)NULL);
if (!result) {
printf("IPC failure: WriteFile failed with %d.\n", GetLastError());
return -1;
}
return 0;
}
int ipcRecvData(ipcHandle *handle, void *data, size_t sz) {
DWORD cbRead = 0;
if (!ReadFile(handle->hMailslot[0], data, (DWORD)sz, &cbRead, NULL)) {
printf("IPC failure: ReadFile failed with %d.\n", GetLastError());
return -1;
}
if (sz != (size_t)cbRead) {
printf(
"IPC failure: ReadFile didn't receive the expected number of bytes\n");
return -1;
}
return 0;
}
int ipcSendShareableHandles(
ipcHandle *handle, const std::vector<ShareableHandle> &shareableHandles,
const std::vector<Process> &processes) {
// Send all shareable handles to every single process.
for (int i = 0; i < processes.size(); i++) {
HANDLE hProcess =
OpenProcess(PROCESS_DUP_HANDLE, FALSE, processes[i].dwProcessId);
if (hProcess == INVALID_HANDLE_VALUE) {
printf("IPC failure: OpenProcess failed (%d)\n", GetLastError());
return -1;
}
for (int j = 0; j < shareableHandles.size(); j++) {
HANDLE hDup = INVALID_HANDLE_VALUE;
// Duplicate the handle into the target process's space
if (!DuplicateHandle(GetCurrentProcess(), shareableHandles[j], hProcess,
&hDup, 0, FALSE, DUPLICATE_SAME_ACCESS)) {
printf("IPC failure: DuplicateHandle failed (%d)\n", GetLastError());
return -1;
}
checkIpcErrors(ipcSendData(handle->hMailslot[i], &hDup, sizeof(hDup)));
}
CloseHandle(hProcess);
}
return 0;
}
int ipcRecvShareableHandles(ipcHandle *handle,
std::vector<ShareableHandle> &shareableHandles) {
for (int i = 0; i < shareableHandles.size(); i++) {
checkIpcErrors(
ipcRecvData(handle, &shareableHandles[i], sizeof(shareableHandles[i])));
}
return 0;
}
int ipcCloseSocket(ipcHandle *handle) {
for (int i = 0; i < handle->hMailslot.size(); i++) {
CloseHandle(handle->hMailslot[i]);
}
delete handle;
return 0;
}
int ipcCloseShareableHandle(ShareableHandle shHandle) {
CloseHandle(shHandle);
return 0;
}
#endif

View File

@@ -0,0 +1,120 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef HELPER_MULTIPROCESS_H
#define HELPER_MULTIPROCESS_H
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
#ifndef WIN32_LEAN_AND_MEAN
#define WIN32_LEAN_AND_MEAN
#endif
#include <windows.h>
#include <iostream>
#include <stdio.h>
#include <tchar.h>
#include <strsafe.h>
#include <sddl.h>
#include <aclapi.h>
#include <winternl.h>
#else
#include <stdio.h>
#include <fcntl.h>
#include <sys/mman.h>
#include <unistd.h>
#include <errno.h>
#include <sys/wait.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <memory.h>
#include <sys/un.h>
#endif
#include <vector>
typedef struct sharedMemoryInfo_st {
void *addr;
size_t size;
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
HANDLE shmHandle;
#else
int shmFd;
#endif
} sharedMemoryInfo;
int sharedMemoryCreate(const char *name, size_t sz, sharedMemoryInfo *info);
int sharedMemoryOpen(const char *name, size_t sz, sharedMemoryInfo *info);
void sharedMemoryClose(sharedMemoryInfo *info);
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
typedef PROCESS_INFORMATION Process;
#else
typedef pid_t Process;
#endif
int spawnProcess(Process *process, const char *app, char * const *args);
int waitProcess(Process *process);
#define checkIpcErrors(ipcFuncResult) \
if (ipcFuncResult == -1) { fprintf(stderr, "Failure at %u %s\n", __LINE__, __FILE__); exit(EXIT_FAILURE); }
#if defined(__linux__)
struct ipcHandle_st {
int socket;
char *socketName;
};
typedef int ShareableHandle;
#elif defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
struct ipcHandle_st {
std::vector<HANDLE> hMailslot; // 1 Handle in case of child and `num children` Handles for parent.
};
typedef HANDLE ShareableHandle;
#endif
typedef struct ipcHandle_st ipcHandle;
int
ipcCreateSocket(ipcHandle *&handle, const char *name, const std::vector<Process>& processes);
int
ipcOpenSocket(ipcHandle *&handle);
int
ipcCloseSocket(ipcHandle *handle);
int
ipcRecvShareableHandles(ipcHandle *handle, std::vector<ShareableHandle>& shareableHandles);
int
ipcSendShareableHandles(ipcHandle *handle, const std::vector<ShareableHandle>& shareableHandles, const std::vector<Process>& processes);
int
ipcCloseShareableHandle(ShareableHandle shHandle);
#endif // HELPER_MULTIPROCESS_H

428
Common/helper_nvJPEG.hxx Normal file
View File

@@ -0,0 +1,428 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
// This sample needs at least CUDA 10.1.
// It demonstrates usages of the nvJPEG library
#ifndef NV_JPEG_EXAMPLE
#define NV_JPEG_EXAMPLE
#ifdef _WIN64
#include <windows.h>
#endif
#include "cuda_runtime.h"
#include "nvjpeg.h"
#include "helper_cuda.h"
#include "helper_timer.h"
#include <cstdlib>
#include <fstream>
#include <iostream>
#include <sstream>
#include <string>
#include <vector>
#include <string.h> // strcmpi
#ifndef _WIN64
#include <sys/time.h> // timings
#include <dirent.h> // linux dir traverse
#include <unistd.h>
#endif
#include <sys/stat.h>
#include <sys/types.h>
// write bmp, input - RGB, device
int writeBMP(const char *filename, const unsigned char *d_chanR, int pitchR,
const unsigned char *d_chanG, int pitchG,
const unsigned char *d_chanB, int pitchB, int width, int height) {
unsigned int headers[13];
FILE *outfile;
int extrabytes;
int paddedsize;
int x;
int y;
int n;
int red, green, blue;
std::vector<unsigned char> vchanR(height * width);
std::vector<unsigned char> vchanG(height * width);
std::vector<unsigned char> vchanB(height * width);
unsigned char *chanR = vchanR.data();
unsigned char *chanG = vchanG.data();
unsigned char *chanB = vchanB.data();
checkCudaErrors(cudaMemcpy2D(chanR, (size_t)width, d_chanR, (size_t)pitchR,
width, height, cudaMemcpyDeviceToHost));
checkCudaErrors(cudaMemcpy2D(chanG, (size_t)width, d_chanG, (size_t)pitchR,
width, height, cudaMemcpyDeviceToHost));
checkCudaErrors(cudaMemcpy2D(chanB, (size_t)width, d_chanB, (size_t)pitchR,
width, height, cudaMemcpyDeviceToHost));
extrabytes =
4 - ((width * 3) % 4); // How many bytes of padding to add to each
// horizontal line - the size of which must
// be a multiple of 4 bytes.
if (extrabytes == 4) extrabytes = 0;
paddedsize = ((width * 3) + extrabytes) * height;
// Headers...
// Note that the "BM" identifier in bytes 0 and 1 is NOT included in these
// "headers".
headers[0] = paddedsize + 54; // bfSize (whole file size)
headers[1] = 0; // bfReserved (both)
headers[2] = 54; // bfOffbits
headers[3] = 40; // biSize
headers[4] = width; // biWidth
headers[5] = height; // biHeight
// Would have biPlanes and biBitCount in position 6, but they're shorts.
// It's easier to write them out separately (see below) than pretend
// they're a single int, especially with endian issues...
headers[7] = 0; // biCompression
headers[8] = paddedsize; // biSizeImage
headers[9] = 0; // biXPelsPerMeter
headers[10] = 0; // biYPelsPerMeter
headers[11] = 0; // biClrUsed
headers[12] = 0; // biClrImportant
if (!(outfile = fopen(filename, "wb"))) {
std::cerr << "Cannot open file: " << filename << std::endl;
return 1;
}
//
// Headers begin...
// When printing ints and shorts, we write out 1 character at a time to avoid
// endian issues.
//
fprintf(outfile, "BM");
for (n = 0; n <= 5; n++) {
fprintf(outfile, "%c", headers[n] & 0x000000FF);
fprintf(outfile, "%c", (headers[n] & 0x0000FF00) >> 8);
fprintf(outfile, "%c", (headers[n] & 0x00FF0000) >> 16);
fprintf(outfile, "%c", (headers[n] & (unsigned int)0xFF000000) >> 24);
}
// These next 4 characters are for the biPlanes and biBitCount fields.
fprintf(outfile, "%c", 1);
fprintf(outfile, "%c", 0);
fprintf(outfile, "%c", 24);
fprintf(outfile, "%c", 0);
for (n = 7; n <= 12; n++) {
fprintf(outfile, "%c", headers[n] & 0x000000FF);
fprintf(outfile, "%c", (headers[n] & 0x0000FF00) >> 8);
fprintf(outfile, "%c", (headers[n] & 0x00FF0000) >> 16);
fprintf(outfile, "%c", (headers[n] & (unsigned int)0xFF000000) >> 24);
}
//
// Headers done, now write the data...
//
for (y = height - 1; y >= 0;
y--) // BMP image format is written from bottom to top...
{
for (x = 0; x <= width - 1; x++) {
red = chanR[y * width + x];
green = chanG[y * width + x];
blue = chanB[y * width + x];
if (red > 255) red = 255;
if (red < 0) red = 0;
if (green > 255) green = 255;
if (green < 0) green = 0;
if (blue > 255) blue = 255;
if (blue < 0) blue = 0;
// Also, it's written in (b,g,r) format...
fprintf(outfile, "%c", blue);
fprintf(outfile, "%c", green);
fprintf(outfile, "%c", red);
}
if (extrabytes) // See above - BMP lines must be of lengths divisible by 4.
{
for (n = 1; n <= extrabytes; n++) {
fprintf(outfile, "%c", 0);
}
}
}
fclose(outfile);
return 0;
}
// write bmp, input - RGB, device
int writeBMPi(const char *filename, const unsigned char *d_RGB, int pitch,
int width, int height) {
unsigned int headers[13];
FILE *outfile;
int extrabytes;
int paddedsize;
int x;
int y;
int n;
int red, green, blue;
std::vector<unsigned char> vchanRGB(height * width * 3);
unsigned char *chanRGB = vchanRGB.data();
checkCudaErrors(cudaMemcpy2D(chanRGB, (size_t)width * 3, d_RGB, (size_t)pitch,
width * 3, height, cudaMemcpyDeviceToHost));
extrabytes =
4 - ((width * 3) % 4); // How many bytes of padding to add to each
// horizontal line - the size of which must
// be a multiple of 4 bytes.
if (extrabytes == 4) extrabytes = 0;
paddedsize = ((width * 3) + extrabytes) * height;
// Headers...
// Note that the "BM" identifier in bytes 0 and 1 is NOT included in these
// "headers".
headers[0] = paddedsize + 54; // bfSize (whole file size)
headers[1] = 0; // bfReserved (both)
headers[2] = 54; // bfOffbits
headers[3] = 40; // biSize
headers[4] = width; // biWidth
headers[5] = height; // biHeight
// Would have biPlanes and biBitCount in position 6, but they're shorts.
// It's easier to write them out separately (see below) than pretend
// they're a single int, especially with endian issues...
headers[7] = 0; // biCompression
headers[8] = paddedsize; // biSizeImage
headers[9] = 0; // biXPelsPerMeter
headers[10] = 0; // biYPelsPerMeter
headers[11] = 0; // biClrUsed
headers[12] = 0; // biClrImportant
if (!(outfile = fopen(filename, "wb"))) {
std::cerr << "Cannot open file: " << filename << std::endl;
return 1;
}
//
// Headers begin...
// When printing ints and shorts, we write out 1 character at a time to avoid
// endian issues.
//
fprintf(outfile, "BM");
for (n = 0; n <= 5; n++) {
fprintf(outfile, "%c", headers[n] & 0x000000FF);
fprintf(outfile, "%c", (headers[n] & 0x0000FF00) >> 8);
fprintf(outfile, "%c", (headers[n] & 0x00FF0000) >> 16);
fprintf(outfile, "%c", (headers[n] & (unsigned int)0xFF000000) >> 24);
}
// These next 4 characters are for the biPlanes and biBitCount fields.
fprintf(outfile, "%c", 1);
fprintf(outfile, "%c", 0);
fprintf(outfile, "%c", 24);
fprintf(outfile, "%c", 0);
for (n = 7; n <= 12; n++) {
fprintf(outfile, "%c", headers[n] & 0x000000FF);
fprintf(outfile, "%c", (headers[n] & 0x0000FF00) >> 8);
fprintf(outfile, "%c", (headers[n] & 0x00FF0000) >> 16);
fprintf(outfile, "%c", (headers[n] & (unsigned int)0xFF000000) >> 24);
}
//
// Headers done, now write the data...
//
for (y = height - 1; y >= 0;
y--) // BMP image format is written from bottom to top...
{
for (x = 0; x <= width - 1; x++) {
red = chanRGB[(y * width + x) * 3];
green = chanRGB[(y * width + x) * 3 + 1];
blue = chanRGB[(y * width + x) * 3 + 2];
if (red > 255) red = 255;
if (red < 0) red = 0;
if (green > 255) green = 255;
if (green < 0) green = 0;
if (blue > 255) blue = 255;
if (blue < 0) blue = 0;
// Also, it's written in (b,g,r) format...
fprintf(outfile, "%c", blue);
fprintf(outfile, "%c", green);
fprintf(outfile, "%c", red);
}
if (extrabytes) // See above - BMP lines must be of lengths divisible by 4.
{
for (n = 1; n <= extrabytes; n++) {
fprintf(outfile, "%c", 0);
}
}
}
fclose(outfile);
return 0;
}
int inputDirExists(const char *pathname) {
struct stat info;
if (stat(pathname, &info) != 0) {
return 0; // Directory does not exists
} else if (info.st_mode & S_IFDIR) {
// is a directory
return 1;
} else {
// is not a directory
return 0;
}
}
int readInput(const std::string &sInputPath,
std::vector<std::string> &filelist) {
int error_code = 1;
#ifndef _WIN64
struct stat s;
if (stat(sInputPath.c_str(), &s) == 0) {
if (s.st_mode & S_IFREG) {
filelist.push_back(sInputPath);
} else if (s.st_mode & S_IFDIR) {
// processing each file in directory
DIR *dir_handle;
struct dirent *dir;
dir_handle = opendir(sInputPath.c_str());
std::vector<std::string> filenames;
if (dir_handle) {
error_code = 0;
while ((dir = readdir(dir_handle)) != NULL) {
std::string sFileName = sInputPath + dir->d_name;
if (inputDirExists(sFileName.c_str()))
{
std::string sname = dir->d_name;
if (sname != "." && sname != "..") {
readInput(sInputPath + sname + "/", filelist);
}
}
else
{
filelist.push_back(sFileName);
}
}
closedir(dir_handle);
} else {
std::cout << "Cannot open input directory: " << sInputPath << std::endl;
return error_code;
}
} else {
std::cout << "Cannot open input: " << sInputPath << std::endl;
return error_code;
}
} else {
std::cout << "Cannot find input path " << sInputPath << std::endl;
return error_code;
}
#else
std::string search_path = sInputPath + "/*.*";
WIN32_FIND_DATA fd;
HANDLE hFind = ::FindFirstFile(search_path.c_str(), &fd);
if (hFind != INVALID_HANDLE_VALUE) {
do {
// read all (real) files in current folder
// , delete '!' read other 2 default folder . and ..
if (!(fd.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)) {
std::string temp(sInputPath + "\\" + fd.cFileName);
filelist.push_back(temp);
}
} while (::FindNextFile(hFind, &fd));
::FindClose(hFind);
} else {
std::cout << "Cannot open input directory: " << sInputPath << std::endl;
return error_code;
}
#endif
return 0;
}
int getInputDir(std::string &input_dir, const char *executable_path) {
int found = 0;
if (executable_path != 0) {
std::string executable_name = std::string(executable_path);
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
// Windows path delimiter
size_t delimiter_pos = executable_name.find_last_of('\\');
executable_name.erase(0, delimiter_pos + 1);
if (executable_name.rfind(".exe") != std::string::npos) {
// we strip .exe, only if the .exe is found
executable_name.resize(executable_name.size() - 4);
}
#else
// Linux & OSX path delimiter
size_t delimiter_pos = executable_name.find_last_of('/');
executable_name.erase(0, delimiter_pos + 1);
#endif
// Search in default paths for input images.
std::string pathname = "";
const char *searchPath[] = {
"./images", "../../../../Samples/<executable_name>/images",
"../../../Samples/<executable_name>/images",
"../../Samples/<executable_name>/images"};
for (unsigned int i = 0; i < sizeof(searchPath) / sizeof(char *); ++i) {
std::string pathname(searchPath[i]);
size_t executable_name_pos = pathname.find("<executable_name>");
// If there is executable_name variable in the searchPath
// replace it with the value
if (executable_name_pos != std::string::npos) {
pathname.replace(executable_name_pos, strlen("<executable_name>"),
executable_name);
}
if (inputDirExists(pathname.c_str())) {
input_dir = pathname + "/";
found = 1;
break;
}
}
}
return found;
}
#endif

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -293,6 +293,9 @@ inline char *sdkFindFilePath(const char *filename,
"../../../../Samples/<executable_name>/data/", // up 4 in tree "../../../../Samples/<executable_name>/data/", // up 4 in tree
"../../../Samples/<executable_name>/data/", // up 3 in tree "../../../Samples/<executable_name>/data/", // up 3 in tree
"../../Samples/<executable_name>/data/", // up 2 in tree "../../Samples/<executable_name>/data/", // up 2 in tree
"../../../../Common/data/", // up 4 in tree
"../../../Common/data/", // up 3 in tree
"../../Common/data/" // up 2 in tree
}; };
// Extract the executable name // Extract the executable name

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions

BIN
Common/lib/x64/freeglut.lib Normal file

Binary file not shown.

BIN
Common/lib/x64/glew64.lib Normal file

Binary file not shown.

View File

@@ -1,4 +1,4 @@
/* Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. /* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
* *
* Redistribution and use in source and binary forms, with or without * Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions * modification, are permitted provided that the following conditions
@@ -47,8 +47,8 @@
} \ } \
} while (0) } while (0)
void compileFileToPTX(char *filename, int argc, char **argv, char **ptxResult, void compileFileToCUBIN(char *filename, int argc, char **argv, char **cubinResult,
size_t *ptxResultSize, int requiresCGheaders) { size_t *cubinResultSize, int requiresCGheaders) {
std::ifstream inputFile(filename, std::ifstream inputFile(filename,
std::ios::in | std::ios::binary | std::ios::ate); std::ios::in | std::ios::binary | std::ios::ate);
@@ -68,7 +68,37 @@ void compileFileToPTX(char *filename, int argc, char **argv, char **ptxResult,
int numCompileOptions = 0; int numCompileOptions = 0;
char *compileParams[1]; char *compileParams[2];
int major = 0, minor = 0;
char deviceName[256];
// Picks the best CUDA device available
CUdevice cuDevice = findCudaDeviceDRV(argc, (const char **)argv);
// get compute capabilities and the devicename
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, cuDevice));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, cuDevice));
{
// Compile cubin for the GPU arch on which are going to run cuda kernel.
std::string compileOptions;
compileOptions = "--gpu-architecture=sm_";
compileParams[numCompileOptions] = reinterpret_cast<char *>(
malloc(sizeof(char) * (compileOptions.length() + 10)));
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
sprintf_s(compileParams[numCompileOptions], sizeof(char) * (compileOptions.length() + 10),
"%s%d%d", compileOptions.c_str(), major, minor);
#else
snprintf(compileParams[numCompileOptions], compileOptions.size() + 10, "%s%d%d",
compileOptions.c_str(), major, minor);
#endif
}
numCompileOptions++;
if (requiresCGheaders) { if (requiresCGheaders) {
std::string compileOptions; std::string compileOptions;
@@ -92,13 +122,13 @@ void compileFileToPTX(char *filename, int argc, char **argv, char **ptxResult,
argv[0]); argv[0]);
} }
compileOptions += path.c_str(); compileOptions += path.c_str();
compileParams[0] = reinterpret_cast<char *>( compileParams[numCompileOptions] = reinterpret_cast<char *>(
malloc(sizeof(char) * (compileOptions.length() + 1))); malloc(sizeof(char) * (compileOptions.length() + 1)));
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64) #if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
sprintf_s(compileParams[0], sizeof(char) * (compileOptions.length() + 1), sprintf_s(compileParams[numCompileOptions], sizeof(char) * (compileOptions.length() + 1),
"%s", compileOptions.c_str()); "%s", compileOptions.c_str());
#else #else
snprintf(compileParams[0], compileOptions.size(), "%s", snprintf(compileParams[numCompileOptions], compileOptions.size(), "%s",
compileOptions.c_str()); compileOptions.c_str());
#endif #endif
numCompileOptions++; numCompileOptions++;
@@ -128,19 +158,20 @@ void compileFileToPTX(char *filename, int argc, char **argv, char **ptxResult,
free(log); free(log);
NVRTC_SAFE_CALL("nvrtcCompileProgram", res); NVRTC_SAFE_CALL("nvrtcCompileProgram", res);
// fetch PTX
size_t ptxSize;
NVRTC_SAFE_CALL("nvrtcGetPTXSize", nvrtcGetPTXSize(prog, &ptxSize));
char *ptx = reinterpret_cast<char *>(malloc(sizeof(char) * ptxSize));
NVRTC_SAFE_CALL("nvrtcGetPTX", nvrtcGetPTX(prog, ptx));
NVRTC_SAFE_CALL("nvrtcDestroyProgram", nvrtcDestroyProgram(&prog));
*ptxResult = ptx;
*ptxResultSize = ptxSize;
if (requiresCGheaders) free(compileParams[0]); size_t codeSize;
NVRTC_SAFE_CALL("nvrtcGetCUBINSize", nvrtcGetCUBINSize(prog, &codeSize));
char *code = new char[codeSize];
NVRTC_SAFE_CALL("nvrtcGetCUBIN", nvrtcGetCUBIN(prog, code));
*cubinResult = code;
*cubinResultSize = codeSize;
for (int i = 0; i < numCompileOptions; i++) {
free(compileParams[i]);
}
} }
CUmodule loadPTX(char *ptx, int argc, char **argv) { CUmodule loadCUBIN(char *cubin, int argc, char **argv) {
CUmodule module; CUmodule module;
CUcontext context; CUcontext context;
int major = 0, minor = 0; int major = 0, minor = 0;
@@ -158,11 +189,10 @@ CUmodule loadPTX(char *ptx, int argc, char **argv) {
printf("> GPU Device has SM %d.%d compute capability\n", major, minor); printf("> GPU Device has SM %d.%d compute capability\n", major, minor);
checkCudaErrors(cuInit(0)); checkCudaErrors(cuInit(0));
checkCudaErrors(cuDeviceGet(&cuDevice, 0));
checkCudaErrors(cuCtxCreate(&context, 0, cuDevice)); checkCudaErrors(cuCtxCreate(&context, 0, cuDevice));
checkCudaErrors(cuModuleLoadDataEx(&module, ptx, 0, 0, 0)); checkCudaErrors(cuModuleLoadData(&module, cubin));
free(ptx); free(cubin);
return module; return module;
} }

View File

@@ -0,0 +1,124 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
////////////////////////////////////////////////////////////////////////////////
//
// Utility funcs to wrap up saving a surface or the back buffer as a PPM file
// In addition, wraps up a threshold comparision of two PPMs.
//
// These functions are designed to be used to implement an automated QA testing for SDK samples.
//
// Author: Bryan Dudash
// Email: sdkfeedback@nvidia.com
//
// Copyright (c) NVIDIA Corporation. All rights reserved.
////////////////////////////////////////////////////////////////////////////////
#include <helper_functions.h>
#include <rendercheck_d3d11.h>
HRESULT CheckRenderD3D11::ActiveRenderTargetToPPM(ID3D11Device *pDevice, const char *zFileName)
{
ID3D11DeviceContext *pDeviceCtxt;
pDevice->GetImmediateContext(&pDeviceCtxt);
ID3D11RenderTargetView *pRTV = NULL;
pDeviceCtxt->OMGetRenderTargets(1,&pRTV,NULL);
ID3D11Resource *pSourceResource = NULL;
pRTV->GetResource(&pSourceResource);
return ResourceToPPM(pDevice,pSourceResource,zFileName);
}
HRESULT CheckRenderD3D11::ResourceToPPM(ID3D11Device *pDevice, ID3D11Resource *pResource, const char *zFileName)
{
ID3D11DeviceContext *pDeviceCtxt;
pDevice->GetImmediateContext(&pDeviceCtxt);
D3D11_RESOURCE_DIMENSION rType;
pResource->GetType(&rType);
if (rType != D3D11_RESOURCE_DIMENSION_TEXTURE2D)
{
printf("SurfaceToPPM: pResource is not a 2D texture! Aborting...\n");
return E_FAIL;
}
ID3D11Texture2D *pSourceTexture = (ID3D11Texture2D *)pResource;
ID3D11Texture2D *pTargetTexture = NULL;
D3D11_TEXTURE2D_DESC desc;
pSourceTexture->GetDesc(&desc);
desc.BindFlags = 0;
desc.CPUAccessFlags = D3D11_CPU_ACCESS_READ;
desc.Usage = D3D11_USAGE_STAGING;
if (FAILED(pDevice->CreateTexture2D(&desc,NULL,&pTargetTexture)))
{
printf("SurfaceToPPM: Unable to create target Texture resoruce! Aborting... \n");
return E_FAIL;
}
pDeviceCtxt->CopyResource(pTargetTexture,pSourceTexture);
D3D11_MAPPED_SUBRESOURCE mappedTex2D;
pDeviceCtxt->Map(pTargetTexture, 0, D3D11_MAP_READ,0,&mappedTex2D);
// Need to convert from dx pitch to pitch=width
unsigned char *pPPMData = new unsigned char[desc.Width*desc.Height*4];
for (unsigned int iHeight = 0; iHeight<desc.Height; iHeight++)
{
memcpy(&(pPPMData[iHeight*desc.Width*4]),(unsigned char *)(mappedTex2D.pData)+iHeight*mappedTex2D.RowPitch,desc.Width*4);
}
pDeviceCtxt->Unmap(pTargetTexture, 0);
// Prepends the PPM header info and bumps byte data afterwards
sdkSavePPM4ub(zFileName, pPPMData, desc.Width, desc.Height);
delete [] pPPMData;
pTargetTexture->Release();
return S_OK;
}
bool CheckRenderD3D11::PPMvsPPM(const char *src_file, const char *ref_file, const char *exec_path,
const float epsilon, const float threshold)
{
char *ref_file_path = sdkFindFilePath(ref_file, exec_path);
if (ref_file_path == NULL)
{
printf("CheckRenderD3D11::PPMvsPPM unable to find <%s> in <%s> Aborting comparison!\n", ref_file, exec_path);
printf(">>> Check info.xml and [project//data] folder <%s> <<<\n", ref_file);
printf("Aborting comparison!\n");
printf(" FAILURE!\n");
return false;
}
return sdkComparePPM(src_file,ref_file_path,epsilon,threshold,true) == true;
}

View File

@@ -0,0 +1,52 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#pragma once
#ifndef _RENDERCHECK_D3D11_H_
#define _RENDERCHECK_D3D11_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <d3d11.h>
class CheckRenderD3D11
{
public:
CheckRenderD3D11() {}
static HRESULT ActiveRenderTargetToPPM(ID3D11Device *pDevice, const char *zFileName);
static HRESULT ResourceToPPM(ID3D11Device *pDevice, ID3D11Resource *pResource, const char *zFileName);
static bool PPMvsPPM(const char *src_file, const char *ref_file, const char *exec_path,
const float epsilon, const float threshold = 0.0f);
};
#endif

View File

@@ -1,4 +1,4 @@
Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
Redistribution and use in source and binary forms, with or without Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions modification, are permitted provided that the following conditions
@@ -23,3 +23,6 @@ PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
For additional information on the license terms, see the CUDA EULA at
https://docs.nvidia.com/cuda/eula/index.html

View File

@@ -1,6 +1,6 @@
############################################################################### ###############################################################################
# #
# Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved. # Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
# #
# Redistribution and use in source and binary forms, with or without # Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions # modification, are permitted provided that the following conditions

183
README.md
View File

@@ -1,11 +1,94 @@
# CUDA Samples # CUDA Samples
Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 9.2](https://developer.nvidia.com/cuda-downloads). Samples for CUDA Developers which demonstrates features in CUDA Toolkit. This version supports [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads).
## Release Notes ## Release Notes
This section describes the release notes for the CUDA Samples on GitHub only. This section describes the release notes for the CUDA Samples on GitHub only.
### CUDA 11.3
* Added `streamOrderedAllocationIPC`. Demonstrates Inter Process Communication using one process per GPU for computation.
* Added `simpleCUBLAS_LU`. Demonstrates batched matrix LU decomposition using cuBLAS API `cublas<t>getrfBatched()`
* Updated `simpleVulkan`. Demonstrates use of timeline semaphore.
* Updated multiple samples to use pinned memory using `cudaMallocHost()`.
### CUDA 11.2
* Added `streamOrderedAllocation`. Demonstrates stream ordered memory allocation on a GPU using cudaMallocAsync and cudaMemPool family of APIs.
* Added `streamOrderedAllocationP2P`. Demonstrates peer-to-peer access of stream ordered memory allocated using cudaMallocAsync and cudaMemPool family of APIs.
* Dropped Visual Studio 2015 support from all the windows supported samples.
* FreeImage is no longer distributed with the CUDA Samples. On Windows, see the [Dependencies](#freeimage) section for more details on how to set up FreeImage. On Linux, it is recommended to install FreeImage with your distribution's package manager.
* All the samples using CUDA Pipeline & Arrive-wait barriers are been updated to use new `cuda::pipeline` and `cuda::barrier` interfaces.
* Updated all the samples to build with parallel build option `--threads` of `nvcc` cuda compiler.
* Added `cudaNvSciNvMedia`. Demonstrates CUDA-NvMedia interop via NvSciBuf/NvSciSync APIs.
* Added `simpleGL`. Demonstrates interoperability between CUDA and OpenGL.
### CUDA 11.1
* Added `watershedSegmentationNPP`. Demonstrates how to use the NPP watershed segmentation function.
* Added `batchedLabelMarkersAndLabelCompressionNPP`. Demonstrates how to use the NPP label markers generation and label compression functions based on a Union Find (UF) algorithm including both single image and batched image versions.
* Dropped Visual Studio 2012, 2013 support from all the windows supported samples.
* Added kernel performing warp aggregated atomic max in multi buckets using cg::labeled_partition & cg::reduce in `warpAggregatedAtomicsCG`.
* Added extended CG shuffle mechanics to `shfl_scan` sample.
* Added `cudaOpenMP`. Demonstrates how to use OpenMP API to write an application for multiple GPUs.
* Added `simpleZeroCopy`. Demonstrates how to use zero copy, kernels can read and write directly to pinned system memory.
### CUDA 11.0
* Added `dmmaTensorCoreGemm`. Demonstrates double precision GEMM computation using the Double precision Warp Matrix Multiply and Accumulate (WMMA) API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `bf16TensorCoreGemm`. Demonstrates __nv_bfloat16 (e8m7) GEMM computation using the __nv_bfloat16 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `tf32TensorCoreGemm`. Demonstrates tf32 (e8m10) GEMM computation using the tf32 WMMA API introduced with CUDA 11 in Ampere chip family tensor cores.
* Added `globalToShmemAsyncCopy`. Demonstrates async copy of data from global to shared memory when on compute capability 8.0 or higher. Also demonstrates arrive-wait barrier for synchronization.
* Added `simpleAWBarrier`. Demonstrates arrive wait barriers.
* Added `simpleAttributes`. Demonstrates the stream attributes that affect L2 locality.
* Added warp aggregated atomic multi bucket increments kernel using labeled_partition cooperative groups in `warpAggregatedAtomicsCG` which can be used on compute capability 7.0 and above GPU architectures.
* Added `binaryPartitionCG`. Demonstrates binary partition cooperative groups and reduction within the thread block.
* Added two new reduction kernels in `reduction` one which demonstrates reduce_add_sync intrinstic supported on compute capability 8.0 and another which uses cooperative_groups::reduce function which does thread_block_tile level reduction introduced from CUDA 11.0.
* Added `cudaCompressibleMemory`. Demonstrates compressible memory allocation using cuMemMap API.
* Added `simpleVulkanMMAP`. Demonstrates Vulkan CUDA Interop via cuMemMap APIs.
* Added `concurrentKernels`. Demonstrates the use of CUDA streams for concurrent execution of several kernels on a GPU.
* Dropped Mac OSX support from all samples.
### CUDA 10.2
* Added `simpleD3D11`. Demonstrates CUDA-D3D11 External Resource Interoperability APIs for updating D3D11 buffers from CUDA and synchronization between D3D11 and CUDA with Keyed Mutexes.
* Added `simpleDrvRuntime`. Demonstrates CUDA Driver and Runtime APIs working together to load fatbinary of a CUDA kernel.
* Added `vectorAddMMAP`. Demonstrates how cuMemMap API allows the user to specify the physical properties of their memory while retaining the contiguous nature of their access.
* Added `memMapIPCDrv`. Demonstrates Inter Process Communication using cuMemMap APIs.
* Added `cudaNvSci`. Demonstrates CUDA-NvSciBuf/NvSciSync Interop.
* Added `jacobiCudaGraphs`. Demonstrates Instantiated CUDA Graph Update with Jacobi Iterative Method using different approaches.
* Added `cuSolverSp_LinearSolver`. Demonstrates cuSolverSP's LU, QR and Cholesky factorization.
* Added `MersenneTwisterGP11213`. Demonstrates the Mersenne Twister random number generator GP11213 in cuRAND.
### CUDA 10.1 Update 2
* Added `vulkanImageCUDA`. Demonstrates how to perform Vulkan image - CUDA Interop.
* Added `nvJPEG_encoder`. Demonstrates encoding of jpeg images using NVJPEG Library.
* Added Windows OS support to `nvJPEG` sample.
* Added `boxFilterNPP`. Demonstrates how to use NPP FilterBox function to perform a box filter.
* Added `cannyEdgeDetectorNPP`. Demonstrates the nppiFilterCannyBorder_8u_C1R Canny Edge Detection image filter function.
### CUDA 10.1 Update 1
* Added `NV12toBGRandResize`. Demonstrates how to convert and resize NV12 frames to BGR planars frames using CUDA in batch.
* Added `EGLStream_CUDA_Interop`. Demonstrates data exchange between CUDA and EGL Streams.
* Added `cuSolverDn_LinearSolver`. Demonstrates cuSolverDN's LU, QR and Cholesky factorization.
* Added support of Visual Studio 2019 to all samples supported on [Windows](#windows-1).
### CUDA 10.1
* Added `immaTensorCoreGemm`. Demonstrates integer GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API for integers employing the Tensor Cores.
* Added `simpleIPC`. Demonstrates Inter Process Communication with one process per GPU for computation.
* Added `nvJPEG`. Demonstrates single and batched decoding of jpeg images using NVJPEG Library.
* Added `bandwidthTest`. It measures the memcopy bandwidth of the GPU and memcpy bandwidth across PCI-e.
* Added `reduction`. Demonstrates several important optimization strategies for Data-Parallel Algorithms like reduction.
* Update all the samples to support CUDA 10.1.
### CUDA 10.0
* Added `simpleCudaGraphs`. Demonstrates CUDA Graphs creation, instantiation and launch using Graphs APIs and Stream Capture APIs.
* Added `conjugateGradientCudaGraphs`. Demonstrates conjugate gradient solver on GPU using CUBLAS and CUSPARSE library calls captured and called using CUDA Graph APIs.
* Added `simpleVulkan`. Demonstrates Vulkan - CUDA Interop.
* Added `simpleD3D12`. Demonstrates DX12 - CUDA Interop.
* Added `UnifiedMemoryPerf`. Demonstrates performance comparision of various memory types involved in system.
* Added `p2pBandwidthLatencyTest`. Demonstrates Peer-To-Peer (P2P) data transfers between pairs of GPUs and computes latency and bandwidth.
* Added `systemWideAtomics`. Demonstrates system wide atomic instructions.
* Added `simpleCUBLASXT`. Demonstrates CUBLAS-XT library which performs GEMM operations over multiple GPUs.
* Added Windows OS support to `conjugateGradientMultiDeviceCG` sample.
* Removed support of Visual Studio 2010 from all samples.
### CUDA 9.2 ### CUDA 9.2
This is the first release of CUDA Samples on GitHub: This is the first release of CUDA Samples on GitHub:
@@ -26,8 +109,8 @@ This is the first release of CUDA Samples on GitHub:
### Prerequisites ### Prerequisites
Download and install the [CUDA Toolkit 9.2](https://developer.nvidia.com/cuda-downloads) for your corresponding platform. Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
For system requirements and installation instructions of cuda toolkit, please refer to the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/), the [Windows Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html), and the [Mac Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html). For system requirements and installation instructions of cuda toolkit, please refer to the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/), and the [Windows Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html).
### Getting the CUDA Samples ### Getting the CUDA Samples
@@ -52,7 +135,6 @@ Each individual sample has its own set of solution files at:
`<CUDA_SAMPLES_REPO>\Samples\<sample_dir>\` `<CUDA_SAMPLES_REPO>\Samples\<sample_dir>\`
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used. To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check [DirectX Dependencies](#directx) section for details."
### Linux ### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make: The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
@@ -79,51 +161,44 @@ The samples makefiles can take advantage of certain options:
$ make HOST_COMPILER=g++ $ make HOST_COMPILER=g++
``` ```
### Mac
The Mac samples are built using makefiles. To use the makefiles, change directory into the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where "A B ..." is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use SMS="50 60".
```
$ make SMS="A B ..."
```
* **HOST_COMPILER=<host_compiler>** - override the default clang host compiler. See the [Mac Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=clang
```
## Samples list ## Samples list
### Samples by OS ### Samples by OS
#### Linux #### Linux
**[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[shfl_scan](./Samples/shfl_scan)** | **[deviceQuery](./Samples/deviceQuery)** | **[bandwidthTest](./Samples/bandwidthTest)** | **[batchedLabelMarkersAndLabelCompressionNPP](./Samples/batchedLabelMarkersAndLabelCompressionNPP)** | **[bf16TensorCoreGemm](./Samples/bf16TensorCoreGemm)** | **[binaryPartitionCG](./Samples/binaryPartitionCG)** |
---|---|---|---| ---|---|---|---|
**[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[conjugateGradientMultiDeviceCG](./Samples/conjugateGradientMultiDeviceCG)** | **[matrixMulDrv](./Samples/matrixMulDrv)** | **[simpleCUFFT](./Samples/simpleCUFFT)** | **[boxFilterNPP](./Samples/boxFilterNPP)** | **[cannyEdgeDetectorNPP](./Samples/cannyEdgeDetectorNPP)** | **[concurrentKernels](./Samples/concurrentKernels)** | **[conjugateGradientCudaGraphs](./Samples/conjugateGradientCudaGraphs)** |
**[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[matrixMul](./Samples/matrixMul)** | **[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[conjugateGradientMultiDeviceCG](./Samples/conjugateGradientMultiDeviceCG)** | **[cudaCompressibleMemory](./Samples/cudaCompressibleMemory)** | **[cudaNvSci](./Samples/cudaNvSci)** |
**[cudaNvSciNvMedia](./Samples/cudaNvSciNvMedia)** | **[cudaOpenMP](./Samples/cudaOpenMP)** | **[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[cuSolverDn_LinearSolver](./Samples/cuSolverDn_LinearSolver)** |
**[cuSolverSp_LinearSolver](./Samples/cuSolverSp_LinearSolver)** | **[deviceQuery](./Samples/deviceQuery)** | **[dmmaTensorCoreGemm](./Samples/dmmaTensorCoreGemm)** | **[EGLStream_CUDA_Interop](./Samples/EGLStream_CUDA_Interop)** |
**[globalToShmemAsyncCopy](./Samples/globalToShmemAsyncCopy)** | **[immaTensorCoreGemm](./Samples/immaTensorCoreGemm)** | **[jacobiCudaGraphs](./Samples/jacobiCudaGraphs)** | **[matrixMul](./Samples/matrixMul)** |
**[matrixMulDrv](./Samples/matrixMulDrv)** | **[memMapIPCDrv](./Samples/memMapIPCDrv)** | **[MersenneTwisterGP11213](./Samples/MersenneTwisterGP11213)** | **[NV12toBGRandResize](./Samples/NV12toBGRandResize)** |
**[nvJPEG](./Samples/nvJPEG)** | **[nvJPEG_encoder](./Samples/nvJPEG_encoder)** | **[p2pBandwidthLatencyTest](./Samples/p2pBandwidthLatencyTest)** | **[reduction](./Samples/reduction)** |
**[shfl_scan](./Samples/shfl_scan)** | **[simpleAttributes](./Samples/simpleAttributes)** | **[simpleAWBarrier](./Samples/simpleAWBarrier)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** |
**[simpleCUBLASXT](./Samples/simpleCUBLASXT)** | **[simpleCUBLAS_LU](./Samples/simpleCUBLAS_LU)** | **[simpleCudaGraphs](./Samples/simpleCudaGraphs)** | **[simpleCUFFT](./Samples/simpleCUFFT)** |
**[simpleDrvRuntime](./Samples/simpleDrvRuntime)** | **[simpleGL](./Samples/simpleGL)** | **[simpleIPC](./Samples/simpleIPC)** | **[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** |
**[simpleVulkan](./Samples/simpleVulkan)** | **[simpleVulkanMMAP](./Samples/simpleVulkanMMAP)** | **[simpleZeroCopy](./Samples/simpleZeroCopy)** | **[streamOrderedAllocation](./Samples/streamOrderedAllocation)** |
**[streamOrderedAllocationIPC](./Samples/streamOrderedAllocationIPC)** | **[streamOrderedAllocationP2P](./Samples/streamOrderedAllocationP2P)** | **[systemWideAtomics](./Samples/systemWideAtomics)** | **[tf32TensorCoreGemm](./Samples/tf32TensorCoreGemm)** |
**[UnifiedMemoryPerf](./Samples/UnifiedMemoryPerf)** | **[vectorAddMMAP](./Samples/vectorAddMMAP)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[vulkanImageCUDA](./Samples/vulkanImageCUDA)** |
**[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[watershedSegmentationNPP](./Samples/watershedSegmentationNPP)** |
#### Windows #### Windows
**[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[shfl_scan](./Samples/shfl_scan)** | **[deviceQuery](./Samples/deviceQuery)** | **[bandwidthTest](./Samples/bandwidthTest)** | **[batchedLabelMarkersAndLabelCompressionNPP](./Samples/batchedLabelMarkersAndLabelCompressionNPP)** | **[bf16TensorCoreGemm](./Samples/bf16TensorCoreGemm)** | **[binaryPartitionCG](./Samples/binaryPartitionCG)** |
---|---|---|---| ---|---|---|---|
**[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[matrixMulDrv](./Samples/matrixMulDrv)** | **[simpleCUFFT](./Samples/simpleCUFFT)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[boxFilterNPP](./Samples/boxFilterNPP)** | **[cannyEdgeDetectorNPP](./Samples/cannyEdgeDetectorNPP)** | **[concurrentKernels](./Samples/concurrentKernels)** | **[conjugateGradientCudaGraphs](./Samples/conjugateGradientCudaGraphs)** |
**[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[matrixMul](./Samples/matrixMul)** | **[conjugateGradientMultiBlockCG](./Samples/conjugateGradientMultiBlockCG)** | **[conjugateGradientMultiDeviceCG](./Samples/conjugateGradientMultiDeviceCG)** | **[cudaCompressibleMemory](./Samples/cudaCompressibleMemory)** | **[cudaOpenMP](./Samples/cudaOpenMP)** |
**[cudaTensorCoreGemm](./Samples/cudaTensorCoreGemm)** | **[cuSolverDn_LinearSolver](./Samples/cuSolverDn_LinearSolver)** | **[cuSolverSp_LinearSolver](./Samples/cuSolverSp_LinearSolver)** | **[deviceQuery](./Samples/deviceQuery)** |
#### Mac OSX **[dmmaTensorCoreGemm](./Samples/dmmaTensorCoreGemm)** | **[globalToShmemAsyncCopy](./Samples/globalToShmemAsyncCopy)** | **[immaTensorCoreGemm](./Samples/immaTensorCoreGemm)** | **[jacobiCudaGraphs](./Samples/jacobiCudaGraphs)** |
**[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[shfl_scan](./Samples/shfl_scan)** | **[deviceQuery](./Samples/deviceQuery)** | **[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[matrixMul](./Samples/matrixMul)** | **[matrixMulDrv](./Samples/matrixMulDrv)** | **[memMapIPCDrv](./Samples/memMapIPCDrv)** | **[MersenneTwisterGP11213](./Samples/MersenneTwisterGP11213)** |
---|---|---|---| **[NV12toBGRandResize](./Samples/NV12toBGRandResize)** | **[nvJPEG](./Samples/nvJPEG)** | **[nvJPEG_encoder](./Samples/nvJPEG_encoder)** | **[p2pBandwidthLatencyTest](./Samples/p2pBandwidthLatencyTest)** |
**[matrixMulDrv](./Samples/matrixMulDrv)** | **[simpleCUFFT](./Samples/simpleCUFFT)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** | **[reduction](./Samples/reduction)** | **[shfl_scan](./Samples/shfl_scan)** | **[simpleAttributes](./Samples/simpleAttributes)** | **[simpleAWBarrier](./Samples/simpleAWBarrier)** |
**[matrixMul](./Samples/matrixMul)** | **[simpleCUBLAS](./Samples/simpleCUBLAS)** | **[simpleCUBLASXT](./Samples/simpleCUBLASXT)** | **[simpleCUBLAS_LU](./Samples/simpleCUBLAS_LU)** | **[simpleCudaGraphs](./Samples/simpleCudaGraphs)** |
**[simpleCUFFT](./Samples/simpleCUFFT)** | **[simpleD3D11](./Samples/simpleD3D11)** | **[simpleD3D12](./Samples/simpleD3D12)** | **[simpleDrvRuntime](./Samples/simpleDrvRuntime)** |
**[simpleGL](./Samples/simpleGL)** | **[simpleIPC](./Samples/simpleIPC)** | **[simpleVoteIntrinsics](./Samples/simpleVoteIntrinsics)** | **[simpleVulkan](./Samples/simpleVulkan)** |
**[simpleVulkanMMAP](./Samples/simpleVulkanMMAP)** | **[simpleZeroCopy](./Samples/simpleZeroCopy)** | **[streamOrderedAllocation](./Samples/streamOrderedAllocation)** | **[streamOrderedAllocationP2P](./Samples/streamOrderedAllocationP2P)** |
**[tf32TensorCoreGemm](./Samples/tf32TensorCoreGemm)** | **[UnifiedMemoryPerf](./Samples/UnifiedMemoryPerf)** | **[vectorAddMMAP](./Samples/vectorAddMMAP)** | **[vectorAdd_nvrtc](./Samples/vectorAdd_nvrtc)** |
**[vulkanImageCUDA](./Samples/vulkanImageCUDA)** | **[warpAggregatedAtomicsCG](./Samples/warpAggregatedAtomicsCG)** | **[watershedSegmentationNPP](./Samples/watershedSegmentationNPP)** |
## Dependencies ## Dependencies
@@ -139,7 +214,9 @@ These third-party dependencies are required by some CUDA samples. If available,
#### FreeImage #### FreeImage
FreeImage is an open source imaging library. FreeImage can usually be installed on Linux using your distribution's package manager system. FreeImage can also be downloaded from the [FreeImage website](http://freeimage.sourceforge.net/). FreeImage is also redistributed with the CUDA Samples. FreeImage is an open source imaging library. FreeImage can usually be installed on Linux using your distribution's package manager system. FreeImage can also be downloaded from the FreeImage website.
To set up FreeImage on a Windows system, extract the FreeImage DLL distribution into the folder `../../Common/FreeImage/Dist/x64` such that it contains the .h, .dll, and .lib files.
#### Message Passing Interface #### Message Passing Interface
@@ -151,7 +228,11 @@ Some samples can only be run on a 64-bit operating system.
#### DirectX #### DirectX
DirectX is a collection of APIs designed to allow development of multimedia applications on Microsoft platforms. For Microsoft platforms, NVIDIA's CUDA Driver supports DirectX. Several CUDA Samples for Windows demonstrates CUDA-DirectX Interoperability, for building such samples one needs to install [Direct X SDK (June 2010 or newer)](http://www.microsoft.com/en-us/download/details.aspx?id=6812) , this is required to be installed on Windows 7, Windows 10 and Windows Server 2008, Other Windows OSes do not need to explicitly install the DirectX SDK. DirectX is a collection of APIs designed to allow development of multimedia applications on Microsoft platforms. For Microsoft platforms, NVIDIA's CUDA Driver supports DirectX. Several CUDA Samples for Windows demonstrates CUDA-DirectX Interoperability, for building such samples one needs to install Microsoft Visual Studio 2012 or higher which provides Microsoft Windows SDK for Windows 8.
#### DirectX12
DirectX 12 is a collection of advanced low-level programming APIs which can reduce driver overhead, designed to allow development of multimedia applications on Microsoft platforms starting with Windows 10 OS onwards. For Microsoft platforms, NVIDIA's CUDA Driver supports DirectX. Few CUDA Samples for Windows demonstrates CUDA-DirectX12 Interoperability, for building such samples one needs to install [Windows 10 SDK or higher](https://developer.microsoft.com/en-us/windows/downloads/windows-10-sdk), with VS 2015 or VS 2017.
#### OpenGL #### OpenGL
@@ -161,6 +242,10 @@ OpenGL is a graphics library used for 2D and 3D rendering. On systems which supp
OpenGL ES is an embedded systems graphics library used for 2D and 3D rendering. On systems which support OpenGL ES, NVIDIA's OpenGL ES implementation is provided with the CUDA Driver. OpenGL ES is an embedded systems graphics library used for 2D and 3D rendering. On systems which support OpenGL ES, NVIDIA's OpenGL ES implementation is provided with the CUDA Driver.
#### Vulkan
Vulkan is a low-overhead, cross-platform 3D graphics and compute API. Vulkan targets high-performance realtime 3D graphics applications such as video games and interactive media across all platforms. On systems which support Vulkan, NVIDIA's Vulkan implementation is provided with the CUDA Driver. For building and running Vulkan applications one needs to install the [Vulkan SDK](https://www.lunarg.com/vulkan-sdk/).
#### OpenMP #### OpenMP
OpenMP is an API for multiprocessing programming. OpenMP can be installed using your Linux distribution's package manager system. It usually comes preinstalled with GCC. It can also be found at the [OpenMP website](http://openmp.org/). OpenMP is an API for multiprocessing programming. OpenMP can be installed using your Linux distribution's package manager system. It usually comes preinstalled with GCC. It can also be found at the [OpenMP website](http://openmp.org/).
@@ -185,6 +270,14 @@ EGLOutput is a set of EGL extensions which allow EGL to render directly to the d
EGLSync is a set of EGL extensions which provides sync objects that are synchronization primitive, representing events whose completion can be tested or waited upon. EGLSync is a set of EGL extensions which provides sync objects that are synchronization primitive, representing events whose completion can be tested or waited upon.
#### NVSCI
NvSci is a set of communication interface libraries out of which CUDA interops with NvSciBuf and NvSciSync. NvSciBuf allows applications to allocate and exchange buffers in memory. NvSciSync allows applications to manage synchronization objects which coordinate when sequences of operations begin and end.
#### NvMedia
NvMedia provides powerful processing of multimedia data for true hardware acceleration across NVIDIA Tegra devices. Applications leverage the NvMedia Application Programming Interface (API) to process the image and video data.
### CUDA Features ### CUDA Features
These CUDA features are needed by some CUDA samples. They are provided by either the CUDA Toolkit or CUDA Driver. Some features may not be available on your system. These CUDA features are needed by some CUDA samples. They are provided by either the CUDA Toolkit or CUDA Driver. Some features may not be available on your system.
@@ -211,7 +304,7 @@ CUBLAS (CUDA Basic Linear Algebra Subroutines) is a GPU-accelerated version of t
#### CUDA Interprocess Communication #### CUDA Interprocess Communication
IPC (Interprocess Communication) allows processes to share device pointers. IPC is only available on Linux x86_64 and ppc64le systems. IPC (Interprocess Communication) allows processes to share device pointers.
#### CUFFT #### CUFFT
@@ -237,6 +330,10 @@ NPP (NVIDIA Performance Primitives) provides GPU-accelerated image, video, and s
NVGRAPH is a GPU-accelerated graph analytics library. NVGRAPH is a GPU-accelerated graph analytics library.
#### NVJPEG
NVJPEG library provides high-performance, GPU accelerated JPEG decoding functionality for image formats commonly used in deep learning and hyperscale multimedia applications.
#### NVRTC #### NVRTC
NVRTC (CUDA RunTime Compilation) is a runtime compilation library for CUDA C++. NVRTC (CUDA RunTime Compilation) is a runtime compilation library for CUDA C++.

View File

@@ -0,0 +1,414 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
UBUNTU = $(shell lsb_release -i -s 2>/dev/null | grep -i ubuntu)
SAMPLE_ENABLED := 1
# This sample is not supported on Mac OSX
ifeq ($(TARGET_OS),darwin)
$(info >>> WARNING - EGLStream_CUDA_Interop is not supported on Mac OSX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
# This sample is not supported on ARMv7
ifeq ($(TARGET_ARCH),armv7l)
$(info >>> WARNING - EGLStream_CUDA_Interop is not supported on ARMv7 - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
# This sample is not supported on android
ifeq ($(TARGET_OS),android)
$(info >>> WARNING - EGLStream_CUDA_Interop is not supported on android - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
# This sample is not supported on QNX
ifeq ($(TARGET_OS),qnx)
$(info >>> WARNING - EGLStream_CUDA_Interop is not supported on QNX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# Makefile include to help find EGL Libraries
include ./findegl.mk
# EGL specific libraries
ifneq ($(TARGET_OS),darwin)
LIBRARIES += -lEGL
endif
ifeq ($(TARGET_OS),darwin)
ALL_LDFLAGS += -Xcompiler -F/Library/Frameworks -Xlinker -framework -Xlinker CUDA
else
ifeq ($(TARGET_ARCH),x86_64)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/lib64/stubs
CUDA_SEARCH_PATH += $(CUDA_PATH)/targets/x86_64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-gnueabihf/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/sbsa-linux/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/armv7-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-linux-androideabi/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ARMv7-linux-QNX/lib/stubs
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/aarch64-qnx/lib/stubs
ifdef TARGET_OVERRIDE
CUDA_SEARCH_PATH := $(CUDA_PATH)/targets/$(TARGET_OVERRIDE)/lib/stubs
endif
endif
ifeq ($(TARGET_ARCH),ppc64le)
CUDA_SEARCH_PATH ?= $(CUDA_PATH)/targets/ppc64le-linux/lib/stubs
endif
ifeq ($(HOST_ARCH),ppc64le)
CUDA_SEARCH_PATH += $(CUDA_PATH)/lib64/stubs
endif
CUDALIB ?= $(shell find -L $(CUDA_SEARCH_PATH) -maxdepth 1 -name libcuda.so 2> /dev/null)
ifeq ("$(CUDALIB)","")
$(info >>> WARNING - libcuda.so not found, CUDA Driver is not installed. Please re-install the driver. <<<)
SAMPLE_ENABLED := 0
else
CUDALIB := $(shell echo $(CUDALIB) | sed "s/ .*//" | sed "s/\/libcuda.so//" )
LIBRARIES += -L$(CUDALIB) -lcuda
endif
endif
ALL_CCFLAGS += --threads 0
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: EGLStream_CUDA_Interop
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
cuda_consumer.o:cuda_consumer.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
cuda_producer.o:cuda_producer.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
eglstrm_common.o:eglstrm_common.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
main.o:main.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
EGLStream_CUDA_Interop: cuda_consumer.o cuda_producer.o eglstrm_common.o main.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./EGLStream_CUDA_Interop
clean:
rm -f EGLStream_CUDA_Interop cuda_consumer.o cuda_producer.o eglstrm_common.o main.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/EGLStream_CUDA_Interop
clobber: clean

View File

@@ -0,0 +1,77 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>EGLStream_CUDA_Interop</name>
<cuda_api_list>
<driver>cuDeviceGet</driver>
<driver>cuDeviceGetAttribute</driver>
<driver>cuDeviceComputeCapability</driver>
<driver>cuDeviceGetCount</driver>
<driver>cuDeviceGetName</driver>
<driver>cuGraphicsResourceGetMappedEglFrame</driver>
<driver>cuEGLStreamConsumerAcquireFrame</driver>
<driver>cuEGLStreamConsumerReleaseFrame</driver>
<driver>cuEGLStreamProducerPresentFrame</driver>
<driver>cuCtxCreate</driver>
<driver>cuMemAlloc</driver>
<driver>cuMemFree</driver>
<driver>cuMemcpy3D</driver>
<driver>cuStreamCreate</driver>
<driver>cuCtxPushCurrent</driver>
<driver>cuCtxPopCurrent</driver>
<driver>cuCtxDestroy</driver>
</cuda_api_list>
<description><![CDATA[Demonstrates data exchange between CUDA and EGL Streams.]]></description>
<devicecompilation>whole</devicecompilation>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../common/inc</path>
</includepaths>
<keyconcepts>
<concept level="basic">EGLStreams Interop</concept>
</keyconcepts>
<keywords>
<keyword>EGL Streams</keyword>
</keywords>
<libraries>
<library os="linux">cuda</library>
<library framework="true" os="macosx">CUDA</library>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>main.cpp</primary_file>
<required_dependencies>
<dependency>EGL</dependency>
</required_dependencies>
<scopes>
<scope>1:CUDA Basic Topics</scope>
<scope>2:Graphics Interop</scope>
</scopes>
<sm-arch>sm35</sm-arch>
<sm-arch>sm37</sm-arch>
<sm-arch>sm50</sm-arch>
<sm-arch>sm52</sm-arch>
<sm-arch>sm60</sm-arch>
<sm-arch>sm61</sm-arch>
<sm-arch>sm70</sm-arch>
<sm-arch>sm72</sm-arch>
<sm-arch>sm75</sm-arch>
<sm-arch>sm80</sm-arch>
<sm-arch>sm86</sm-arch>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<arch>aarch64</arch>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>EGLStream CUDA Interop</title>
<type>exe</type>
</entry>

View File

@@ -0,0 +1,64 @@
# EGLStream_CUDA_Interop - EGLStream CUDA Interop
## Description
Demonstrates data exchange between CUDA and EGL Streams.
## Key Concepts
EGLStreams Interop
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux
## Supported CPU Architecture
x86_64, aarch64
## CUDA APIs involved
### [CUDA Driver API](http://docs.nvidia.com/cuda/cuda-driver-api/index.html)
cuDeviceGet, cuDeviceGetAttribute, cuDeviceComputeCapability, cuDeviceGetCount, cuDeviceGetName, cuGraphicsResourceGetMappedEglFrame, cuEGLStreamConsumerAcquireFrame, cuEGLStreamConsumerReleaseFrame, cuEGLStreamProducerPresentFrame, cuCtxCreate, cuMemAlloc, cuMemFree, cuMemcpy3D, cuStreamCreate, cuCtxPushCurrent, cuCtxPopCurrent, cuCtxDestroy
## Dependencies needed to build/run
[EGL](../../README.md#egl)
## Prerequisites
Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, aarch64.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=aarch64` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,318 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Simple CUDA consumer rendering sample app
//
#include "cuda_consumer.h"
#include <helper_cuda_drvapi.h>
#include "eglstrm_common.h"
#if defined(EXTENSION_LIST)
EXTENSION_LIST(EXTLST_EXTERN)
#endif
int checkbuf(FILE *fp1, FILE *fp2);
CUresult cudaConsumerTest(test_cuda_consumer_s *data, char *fileName) {
CUresult cuStatus = CUDA_SUCCESS;
CUarray cudaArr = NULL;
CUeglFrame cudaEgl;
CUgraphicsResource cudaResource;
unsigned int i;
int check_result;
FILE *pInFile1 = NULL, *pInFile2 = NULL, *file_p = NULL;
EGLint streamState = 0;
if (!data) {
printf("%s: Bad parameter\n", __func__);
goto done;
}
if (!eglQueryStreamKHR(g_display, eglStream, EGL_STREAM_STATE_KHR,
&streamState)) {
printf("Cuda consumer, eglQueryStreamKHR EGL_STREAM_STATE_KHR failed\n");
}
if (streamState == EGL_STREAM_STATE_DISCONNECTED_KHR) {
printf("CUDA Consumer: - EGL_STREAM_STATE_DISCONNECTED_KHR received\n");
}
if (streamState == EGL_STREAM_STATE_NEW_FRAME_AVAILABLE_KHR) {
cuStatus = cuEGLStreamConsumerAcquireFrame(&(data->cudaConn), &cudaResource,
NULL, 16000);
if (cuStatus == CUDA_SUCCESS) {
CUdeviceptr pDevPtr = 0;
int bufferSize;
unsigned char *pCudaCopyMem = NULL;
unsigned int copyWidthInBytes = 0, copyHeight = 0;
file_p = fopen(fileName, "wb+");
if (!file_p) {
printf("WriteFrame: file open failed %s\n", fileName);
cuStatus = CUDA_ERROR_UNKNOWN;
goto done;
}
cuStatus =
cuGraphicsResourceGetMappedEglFrame(&cudaEgl, cudaResource, 0, 0);
if (cuStatus != CUDA_SUCCESS) {
printf("Cuda get resource failed with %d\n", cuStatus);
goto done;
}
cuStatus = cuCtxSynchronize();
if (cuStatus != CUDA_SUCCESS) {
printf("cuCtxSynchronize failed \n");
goto done;
}
if (!(cudaEgl.planeCount >= 1 && cudaEgl.planeCount <= 3)) {
printf("Plane count is invalid\nExiting\n");
goto done;
}
for (i = 0; i < cudaEgl.planeCount; i++) {
if (cudaEgl.frameType == CU_EGL_FRAME_TYPE_PITCH) {
pDevPtr = (CUdeviceptr)cudaEgl.frame.pPitch[i];
if (cudaEgl.planeCount == 1) {
bufferSize = cudaEgl.pitch * cudaEgl.height;
copyWidthInBytes = cudaEgl.pitch;
copyHeight = data->height;
} else if (i == 1 && cudaEgl.planeCount == 2) { // YUV 420
// semi-planar
bufferSize = cudaEgl.pitch * cudaEgl.height / 2;
copyWidthInBytes = cudaEgl.pitch;
copyHeight = data->height / 2;
} else {
bufferSize = data->width * data->height;
copyWidthInBytes = data->width;
copyHeight = data->height;
if (i > 0) {
bufferSize >>= 2;
copyWidthInBytes >>= 1;
copyHeight >>= 1;
}
}
} else {
cudaArr = cudaEgl.frame.pArray[i];
if (cudaEgl.planeCount == 1) {
bufferSize = data->width * data->height * 4;
copyWidthInBytes = data->width * 4;
copyHeight = data->height;
} else if (i == 1 && cudaEgl.planeCount == 2) { // YUV 420
// semi-planar
bufferSize = data->width * data->height / 2;
copyWidthInBytes = data->width;
copyHeight = data->height / 2;
} else {
bufferSize = data->width * data->height;
copyWidthInBytes = data->width;
copyHeight = data->height;
if (i > 0) {
bufferSize >>= 2;
copyWidthInBytes >>= 1;
copyHeight >>= 1;
}
}
}
if (i == 0) {
pCudaCopyMem = (unsigned char *)malloc(bufferSize);
if (pCudaCopyMem == NULL) {
printf("pCudaCopyMem malloc failed\n");
goto done;
}
}
memset(pCudaCopyMem, 0, bufferSize);
if (data->pitchLinearOutput) {
cuStatus = cuMemcpyDtoH(pCudaCopyMem, pDevPtr, bufferSize);
if (cuStatus != CUDA_SUCCESS) {
printf(
"cuda_consumer: pitch linear Memcpy failed, bufferSize =%d\n",
bufferSize);
goto done;
}
cuStatus = cuCtxSynchronize();
if (cuStatus != CUDA_SUCCESS) {
printf("cuda_consumer: cuCtxSynchronize failed after memcpy \n");
goto done;
}
} else {
CUDA_MEMCPY3D cpdesc;
memset(&cpdesc, 0, sizeof(cpdesc));
cpdesc.srcXInBytes = cpdesc.srcY = cpdesc.srcZ = cpdesc.srcLOD = 0;
cpdesc.srcMemoryType = CU_MEMORYTYPE_ARRAY;
cpdesc.srcArray = cudaArr;
cpdesc.dstXInBytes = cpdesc.dstY = cpdesc.dstZ = cpdesc.dstLOD = 0;
cpdesc.dstMemoryType = CU_MEMORYTYPE_HOST;
cpdesc.dstHost = (void *)pCudaCopyMem;
cpdesc.WidthInBytes = copyWidthInBytes; // data->width * 4;
cpdesc.Height = copyHeight; // data->height;
cpdesc.Depth = 1;
cuStatus = cuMemcpy3D(&cpdesc);
if (cuStatus != CUDA_SUCCESS) {
printf(
"Cuda consumer: cuMemCpy3D failed, copyWidthInBytes=%d, "
"copyHight=%d\n",
copyWidthInBytes, copyHeight);
}
cuStatus = cuCtxSynchronize();
if (cuStatus != CUDA_SUCCESS) {
printf("cuCtxSynchronize failed after memcpy \n");
}
}
if (cuStatus == CUDA_SUCCESS) {
if (fwrite(pCudaCopyMem, bufferSize, 1, file_p) != 1) {
printf("Cuda consumer: output file write failed\n");
cuStatus = CUDA_ERROR_UNKNOWN;
goto done;
}
}
}
pInFile1 = fopen(data->fileName1, "rb");
if (!pInFile1) {
printf("Failed to open file :%s\n", data->fileName1);
goto done;
}
pInFile2 = fopen(data->fileName2, "rb");
if (!pInFile2) {
printf("Failed to open file :%s\n", data->fileName2);
goto done;
}
rewind(file_p);
check_result = checkbuf(file_p, pInFile1);
if (check_result == -1) {
rewind(file_p);
check_result = checkbuf(file_p, pInFile2);
if (check_result == -1) {
printf("Frame received does not match any valid image: FAILED\n");
} else {
printf("Frame check Passed\n");
}
} else {
printf("Frame check Passed\n");
}
if (pCudaCopyMem) {
free(pCudaCopyMem);
pCudaCopyMem = NULL;
}
cuStatus =
cuEGLStreamConsumerReleaseFrame(&data->cudaConn, cudaResource, NULL);
if (cuStatus != CUDA_SUCCESS) {
printf("cuEGLStreamConsumerReleaseFrame failed with cuStatus = %d\n",
cuStatus);
goto done;
}
} else {
printf("cuda AcquireFrame FAILED with cuStatus=%d\n", cuStatus);
goto done;
}
}
done:
if (file_p) {
fclose(file_p);
file_p = NULL;
}
if (pInFile1) {
fclose(pInFile1);
pInFile1 = NULL;
}
if (pInFile1) {
fclose(pInFile2);
pInFile2 = NULL;
}
return cuStatus;
}
int checkbuf(FILE *fp1, FILE *fp2) {
int match = 0;
int ch1, ch2;
if (fp1 == NULL) {
printf("Invalid file pointer for first file\n");
return -1;
} else if (fp2 == NULL) {
printf("Invalid file pointer for second file\n");
return -1;
} else {
ch1 = getc(fp1);
ch2 = getc(fp2);
while ((ch1 != EOF) && (ch2 != EOF) && (ch1 == ch2)) {
ch1 = getc(fp1);
ch2 = getc(fp2);
}
if (ch1 == ch2) {
match = 1;
} else if (ch1 != ch2) {
match = -1;
}
}
return match;
}
CUresult cudaDeviceCreateConsumer(test_cuda_consumer_s *cudaConsumer,
CUdevice device) {
CUresult status = CUDA_SUCCESS;
if (CUDA_SUCCESS != (status = cuInit(0))) {
printf("Failed to initialize CUDA\n");
return status;
}
int major = 0, minor = 0;
char deviceName[256];
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, device));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, device));
checkCudaErrors(cuDeviceGetName(deviceName, 256, device));
printf(
"CUDA Consumer on GPU Device %d: \"%s\" with compute capability "
"%d.%d\n\n",
device, deviceName, major, minor);
if (CUDA_SUCCESS !=
(status = cuCtxCreate(&cudaConsumer->context, 0, device))) {
printf("failed to create CUDA context\n");
return status;
}
checkCudaErrors(cuCtxPopCurrent(&cudaConsumer->context));
return status;
}
void cuda_consumer_init(test_cuda_consumer_s *cudaConsumer, TestArgs *args) {
cudaConsumer->pitchLinearOutput = args->pitchLinearOutput;
cudaConsumer->width = args->inputWidth;
cudaConsumer->height = args->inputHeight;
cudaConsumer->fileName1 = args->infile1;
cudaConsumer->fileName2 = args->infile2;
cudaConsumer->outFile1 = "cuda_out1.yuv";
cudaConsumer->outFile2 = "cuda_out2.yuv";
}
CUresult cuda_consumer_deinit(test_cuda_consumer_s *cudaConsumer) {
return cuEGLStreamConsumerDisconnect(&cudaConsumer->cudaConn);
}

View File

@@ -0,0 +1,62 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: CUDA consumer header file
//
#ifndef _CUDA_CONSUMER_H_
#define _CUDA_CONSUMER_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "cudaEGL.h"
#include "eglstrm_common.h"
extern EGLStreamKHR eglStream;
extern EGLDisplay g_display;
typedef struct _test_cuda_consumer_s {
CUcontext context;
CUeglStreamConnection cudaConn;
bool pitchLinearOutput;
unsigned int width;
unsigned int height;
char *fileName1;
char *fileName2;
char *outFile1;
char *outFile2;
unsigned int frameCount;
} test_cuda_consumer_s;
void cuda_consumer_init(test_cuda_consumer_s *cudaConsumer, TestArgs *args);
CUresult cuda_consumer_deinit(test_cuda_consumer_s *cudaConsumer);
CUresult cudaConsumerTest(test_cuda_consumer_s *data, char *outFile);
CUresult cudaDeviceCreateConsumer(test_cuda_consumer_s *cudaConsumer,
CUdevice device);
#endif

Binary file not shown.

Binary file not shown.

View File

@@ -0,0 +1,406 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Simple cuda EGL stream producer app
//
#include "cuda_producer.h"
#include <helper_cuda_drvapi.h>
#include "cudaEGL.h"
#include "eglstrm_common.h"
#if defined(EXTENSION_LIST)
EXTENSION_LIST(EXTLST_EXTERN)
#endif
static CUresult cudaProducerReadYUVFrame(FILE *file, unsigned int frameNum,
unsigned int width,
unsigned int height,
unsigned char *pBuff) {
int bOrderUV = 0;
unsigned char *pYBuff, *pUBuff, *pVBuff, *pChroma;
unsigned int frameSize = (width * height * 3) / 2;
CUresult ret = CUDA_SUCCESS;
unsigned int i;
if (!pBuff || !file) return CUDA_ERROR_FILE_NOT_FOUND;
pYBuff = pBuff;
// YVU order in the buffer
pVBuff = pYBuff + width * height;
pUBuff = pVBuff + width * height / 4;
if (fseek(file, frameNum * frameSize, SEEK_SET)) {
printf("ReadYUVFrame: Error seeking file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
// read Y U V separately
for (i = 0; i < height; i++) {
if (fread(pYBuff, width, 1, file) != 1) {
printf("ReadYUVFrame: Error reading file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
pYBuff += width;
}
pChroma = bOrderUV ? pUBuff : pVBuff;
for (i = 0; i < height / 2; i++) {
if (fread(pChroma, width / 2, 1, file) != 1) {
printf("ReadYUVFrame: Error reading file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
pChroma += width / 2;
}
pChroma = bOrderUV ? pVBuff : pUBuff;
for (i = 0; i < height / 2; i++) {
if (fread(pChroma, width / 2, 1, file) != 1) {
printf("ReadYUVFrame: Error reading file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
pChroma += width / 2;
}
done:
return ret;
}
static CUresult cudaProducerReadARGBFrame(FILE *file, unsigned int frameNum,
unsigned int width,
unsigned int height,
unsigned char *pBuff) {
unsigned int frameSize = width * height * 4;
CUresult ret = CUDA_SUCCESS;
if (!pBuff || !file) return CUDA_ERROR_FILE_NOT_FOUND;
if (fseek(file, frameNum * frameSize, SEEK_SET)) {
printf("ReadYUVFrame: Error seeking file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
// read ARGB data
if (fread(pBuff, frameSize, 1, file) != 1) {
if (feof(file))
printf("ReadARGBFrame: file read to the end\n");
else
printf("ReadARGBFrame: Error reading file: %p\n", file);
ret = CUDA_ERROR_NOT_PERMITTED;
goto done;
}
done:
return ret;
}
CUresult cudaProducerTest(test_cuda_producer_s *cudaProducer, char *file) {
int framenum = 0;
CUarray cudaArr[3] = {0};
CUdeviceptr cudaPtr[3] = {0, 0, 0};
unsigned int bufferSize;
CUresult cuStatus = CUDA_SUCCESS;
unsigned int i, surfNum, uvOffset[3] = {0};
unsigned int copyWidthInBytes[3] = {0, 0, 0}, copyHeight[3] = {0, 0, 0};
CUeglColorFormat eglColorFormat;
FILE *file_p;
CUeglFrame cudaEgl;
CUcontext oldContext;
file_p = fopen(file, "rb");
if (!file_p) {
printf("CudaProducer: Error opening file: %s\n", file);
goto done;
}
if (cudaProducer->pitchLinearOutput) {
if (cudaProducer->isARGB) {
cudaPtr[0] = cudaProducer->cudaPtrARGB[0];
} else { // YUV case
for (i = 0; i < 3; i++) {
if (i == 0) {
bufferSize = cudaProducer->width * cudaProducer->height;
} else {
bufferSize = cudaProducer->width * cudaProducer->height / 4;
}
cudaPtr[i] = cudaProducer->cudaPtrYUV[i];
}
}
} else {
if (cudaProducer->isARGB) {
cudaArr[0] = cudaProducer->cudaArrARGB[0];
} else {
for (i = 0; i < 3; i++) {
cudaArr[i] = cudaProducer->cudaArrYUV[i];
}
}
}
uvOffset[0] = 0;
if (cudaProducer->isARGB) {
if (CUDA_SUCCESS !=
cudaProducerReadARGBFrame(file_p, framenum, cudaProducer->width,
cudaProducer->height, cudaProducer->pBuff)) {
printf("cuda producer, read ARGB frame failed\n");
goto done;
}
copyWidthInBytes[0] = cudaProducer->width * 4;
copyHeight[0] = cudaProducer->height;
surfNum = 1;
eglColorFormat = CU_EGL_COLOR_FORMAT_ARGB;
} else {
if (CUDA_SUCCESS !=
cudaProducerReadYUVFrame(file_p, framenum, cudaProducer->width,
cudaProducer->height, cudaProducer->pBuff)) {
printf("cuda producer, reading YUV frame failed\n");
goto done;
}
surfNum = 3;
eglColorFormat = CU_EGL_COLOR_FORMAT_YUV420_PLANAR;
copyWidthInBytes[0] = cudaProducer->width;
copyHeight[0] = cudaProducer->height;
copyWidthInBytes[1] = cudaProducer->width / 2;
copyHeight[1] = cudaProducer->height / 2;
copyWidthInBytes[2] = cudaProducer->width / 2;
copyHeight[2] = cudaProducer->height / 2;
uvOffset[1] = cudaProducer->width * cudaProducer->height;
uvOffset[2] =
uvOffset[1] + cudaProducer->width / 2 * cudaProducer->height / 2;
}
if (cudaProducer->pitchLinearOutput) {
for (i = 0; i < surfNum; i++) {
cuStatus =
cuMemcpy(cudaPtr[i], (CUdeviceptr)(cudaProducer->pBuff + uvOffset[i]),
copyWidthInBytes[i] * copyHeight[i]);
if (cuStatus != CUDA_SUCCESS) {
printf("Cuda producer: cuMemCpy pitchlinear failed, cuStatus =%d\n",
cuStatus);
goto done;
}
}
} else {
// copy cudaProducer->pBuff to cudaArray
CUDA_MEMCPY3D cpdesc;
for (i = 0; i < surfNum; i++) {
memset(&cpdesc, 0, sizeof(cpdesc));
cpdesc.srcXInBytes = cpdesc.srcY = cpdesc.srcZ = cpdesc.srcLOD = 0;
cpdesc.srcMemoryType = CU_MEMORYTYPE_HOST;
cpdesc.srcHost = (void *)(cudaProducer->pBuff + uvOffset[i]);
cpdesc.dstXInBytes = cpdesc.dstY = cpdesc.dstZ = cpdesc.dstLOD = 0;
cpdesc.dstMemoryType = CU_MEMORYTYPE_ARRAY;
cpdesc.dstArray = cudaArr[i];
cpdesc.WidthInBytes = copyWidthInBytes[i];
cpdesc.Height = copyHeight[i];
cpdesc.Depth = 1;
cuStatus = cuMemcpy3D(&cpdesc);
if (cuStatus != CUDA_SUCCESS) {
printf("Cuda producer: cuMemCpy failed, cuStatus =%d\n", cuStatus);
goto done;
}
}
}
for (i = 0; i < surfNum; i++) {
if (cudaProducer->pitchLinearOutput)
cudaEgl.frame.pPitch[i] = (void *)cudaPtr[i];
else
cudaEgl.frame.pArray[i] = cudaArr[i];
}
cudaEgl.width = copyWidthInBytes[0];
cudaEgl.depth = 1;
cudaEgl.height = copyHeight[0];
cudaEgl.pitch = cudaProducer->pitchLinearOutput ? cudaEgl.width : 0;
cudaEgl.frameType = cudaProducer->pitchLinearOutput ? CU_EGL_FRAME_TYPE_PITCH
: CU_EGL_FRAME_TYPE_ARRAY;
cudaEgl.planeCount = surfNum;
cudaEgl.numChannels = (eglColorFormat == CU_EGL_COLOR_FORMAT_ARGB) ? 4 : 1;
cudaEgl.eglColorFormat = eglColorFormat;
cudaEgl.cuFormat = CU_AD_FORMAT_UNSIGNED_INT8;
static int numFramesPresented = 0;
// If there is a frame presented before we check if consumer
// is done with it using cuEGLStreamProducerReturnFrame.
while (numFramesPresented) {
CUeglFrame returnedCudaEgl;
cuStatus = cuEGLStreamProducerReturnFrame(&cudaProducer->cudaConn,
&returnedCudaEgl, NULL);
if (cuStatus == CUDA_ERROR_LAUNCH_TIMEOUT) {
continue;
} else if (cuStatus != CUDA_SUCCESS) {
printf("cuda Producer return frame FAILED with custatus= %d\n", cuStatus);
return cuStatus;
} else {
numFramesPresented--;
}
}
cuStatus =
cuEGLStreamProducerPresentFrame(&cudaProducer->cudaConn, cudaEgl, NULL);
if (cuStatus != CUDA_SUCCESS) {
printf("cuda Producer present frame FAILED with custatus= %d\n", cuStatus);
goto done;
}
numFramesPresented++;
done:
if (file_p) {
fclose(file_p);
file_p = NULL;
}
return cuStatus;
}
CUresult cudaDeviceCreateProducer(test_cuda_producer_s *cudaProducer,
CUdevice device) {
CUresult status = CUDA_SUCCESS;
if (CUDA_SUCCESS != (status = cuInit(0))) {
printf("Failed to initialize CUDA\n");
return status;
}
int major = 0, minor = 0;
char deviceName[256];
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, device));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, device));
checkCudaErrors(cuDeviceGetName(deviceName, 256, device));
printf(
"CUDA Producer on GPU Device %d: \"%s\" with compute capability "
"%d.%d\n\n",
device, deviceName, major, minor);
if (major < 6) {
printf(
"EGLStreams_CUDA_Interop requires SM 6.0 or higher arch GPU. "
"Exiting...\n");
exit(2); // EXIT_WAIVED
}
if (CUDA_SUCCESS !=
(status = cuCtxCreate(&cudaProducer->context, 0, device))) {
printf("failed to create CUDA context\n");
return status;
}
status = cuMemAlloc(&cudaProducer->cudaPtrARGB[0], (WIDTH * HEIGHT * 4));
if (status != CUDA_SUCCESS) {
printf("Create CUDA pointer failed, cuStatus=%d\n", status);
return status;
}
status = cuMemAlloc(&cudaProducer->cudaPtrYUV[0], (WIDTH * HEIGHT));
if (status != CUDA_SUCCESS) {
printf("Create CUDA pointer failed, cuStatus=%d\n", status);
return status;
}
status = cuMemAlloc(&cudaProducer->cudaPtrYUV[1], (WIDTH * HEIGHT) / 4);
if (status != CUDA_SUCCESS) {
printf("Create CUDA pointer failed, cuStatus=%d\n", status);
return status;
}
status = cuMemAlloc(&cudaProducer->cudaPtrYUV[2], (WIDTH * HEIGHT) / 4);
if (status != CUDA_SUCCESS) {
printf("Create CUDA pointer failed, cuStatus=%d\n", status);
return status;
}
CUDA_ARRAY3D_DESCRIPTOR desc = {0};
desc.Format = CU_AD_FORMAT_UNSIGNED_INT8;
desc.Depth = 1;
desc.Flags = CUDA_ARRAY3D_SURFACE_LDST;
desc.NumChannels = 4;
desc.Width = WIDTH * 4;
desc.Height = HEIGHT;
status = cuArray3DCreate(&cudaProducer->cudaArrARGB[0], &desc);
if (status != CUDA_SUCCESS) {
printf("Create CUDA array failed, cuStatus=%d\n", status);
return status;
}
for (int i = 0; i < 3; i++) {
if (i == 0) {
desc.NumChannels = 1;
desc.Width = WIDTH;
desc.Height = HEIGHT;
} else { // U/V surface as planar
desc.NumChannels = 1;
desc.Width = WIDTH / 2;
desc.Height = HEIGHT / 2;
}
status = cuArray3DCreate(&cudaProducer->cudaArrYUV[i], &desc);
if (status != CUDA_SUCCESS) {
printf("Create CUDA array failed, cuStatus=%d\n", status);
return status;
}
}
cudaProducer->pBuff = (unsigned char *)malloc((WIDTH * HEIGHT * 4));
if (!cudaProducer->pBuff) {
printf("CudaProducer: Failed to allocate image buffer\n");
}
checkCudaErrors(cuCtxPopCurrent(&cudaProducer->context));
return status;
}
void cudaProducerInit(test_cuda_producer_s *cudaProducer, EGLDisplay eglDisplay,
EGLStreamKHR eglStream, TestArgs *args) {
cudaProducer->fileName1 = args->infile1;
cudaProducer->fileName2 = args->infile2;
cudaProducer->frameCount = 2;
cudaProducer->width = args->inputWidth;
cudaProducer->height = args->inputHeight;
cudaProducer->isARGB = args->isARGB;
cudaProducer->pitchLinearOutput = args->pitchLinearOutput;
// Set cudaProducer default parameters
cudaProducer->eglDisplay = eglDisplay;
cudaProducer->eglStream = eglStream;
}
CUresult cudaProducerDeinit(test_cuda_producer_s *cudaProducer) {
if (cudaProducer->pBuff) free(cudaProducer->pBuff);
checkCudaErrors(cuMemFree(cudaProducer->cudaPtrARGB[0]));
checkCudaErrors(cuMemFree(cudaProducer->cudaPtrYUV[0]));
checkCudaErrors(cuMemFree(cudaProducer->cudaPtrYUV[1]));
checkCudaErrors(cuMemFree(cudaProducer->cudaPtrYUV[2]));
checkCudaErrors(cuArrayDestroy(cudaProducer->cudaArrARGB[0]));
checkCudaErrors(cuArrayDestroy(cudaProducer->cudaArrYUV[0]));
checkCudaErrors(cuArrayDestroy(cudaProducer->cudaArrYUV[1]));
checkCudaErrors(cuArrayDestroy(cudaProducer->cudaArrYUV[2]));
return cuEGLStreamProducerDisconnect(&cudaProducer->cudaConn);
}

View File

@@ -0,0 +1,68 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Simple cuda producer header file
//
#ifndef _CUDA_PRODUCER_H_
#define _CUDA_PRODUCER_H_
#include <EGL/egl.h>
#include <EGL/eglext.h>
#include "cudaEGL.h"
#include "eglstrm_common.h"
extern EGLStreamKHR eglStream;
extern EGLDisplay g_display;
typedef struct _test_cuda_producer_s {
// Stream params
char *fileName1;
char *fileName2;
unsigned char *pBuff;
int frameCount;
bool isARGB;
bool pitchLinearOutput;
unsigned int width;
unsigned int height;
CUcontext context;
CUeglStreamConnection cudaConn;
CUdeviceptr cudaPtrARGB[1];
CUdeviceptr cudaPtrYUV[3];
CUarray cudaArrARGB[1];
CUarray cudaArrYUV[3];
EGLStreamKHR eglStream;
EGLDisplay eglDisplay;
} test_cuda_producer_s;
void cudaProducerInit(test_cuda_producer_s *cudaProducer, EGLDisplay eglDisplay,
EGLStreamKHR eglStream, TestArgs *args);
CUresult cudaProducerTest(test_cuda_producer_s *parserArg, char *file);
CUresult cudaProducerDeinit(test_cuda_producer_s *cudaProducer);
CUresult cudaDeviceCreateProducer(test_cuda_producer_s *cudaProducer,
CUdevice device);
#endif

Binary file not shown.

Binary file not shown.

View File

@@ -0,0 +1,139 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Common egl stream functions
//
#include "eglstrm_common.h"
EGLStreamKHR eglStream;
EGLDisplay g_display;
EGLAttrib cudaIndex;
#if defined(EXTENSION_LIST)
EXTENSION_LIST(EXTLST_DECL)
typedef void (*extlst_fnptr_t)(void);
static struct {
extlst_fnptr_t *fnptr;
char const *name;
} extensionList[] = {EXTENSION_LIST(EXTLST_ENTRY)};
int eglSetupExtensions(void) {
unsigned int i;
for (i = 0; i < (sizeof(extensionList) / sizeof(*extensionList)); i++) {
*extensionList[i].fnptr = eglGetProcAddress(extensionList[i].name);
if (*extensionList[i].fnptr == NULL) {
printf("Couldn't get address of %s()\n", extensionList[i].name);
return 0;
}
}
return 1;
}
int EGLStreamInit(int *cuda_device) {
static const EGLint streamAttrMailboxMode[] = {EGL_SUPPORT_REUSE_NV,
EGL_FALSE, EGL_NONE};
EGLBoolean eglStatus;
#define MAX_EGL_DEVICES 4
EGLint numDevices = 0;
EGLDeviceEXT devices[MAX_EGL_DEVICES];
eglStatus = eglQueryDevicesEXT(MAX_EGL_DEVICES, devices, &numDevices);
if (eglStatus != EGL_TRUE) {
printf("Error querying EGL devices\n");
exit(EXIT_FAILURE);
}
if (numDevices == 0) {
printf("No EGL devices found.. Waiving\n");
eglStatus = EGL_FALSE;
exit(EXIT_WAIVED);
}
int egl_device_id = 0;
for (egl_device_id = 0; egl_device_id < numDevices; egl_device_id++) {
eglStatus = eglQueryDeviceAttribEXT(devices[egl_device_id],
EGL_CUDA_DEVICE_NV, &cudaIndex);
if (eglStatus == EGL_TRUE) {
*cuda_device = cudaIndex; // We select first EGL-CUDA Capable device.
printf("Found EGL-CUDA Capable device with CUDA Device id = %d\n",
(int)cudaIndex);
break;
}
}
if (egl_device_id >= numDevices) {
printf("No CUDA Capable EGL Device found.. Waiving execution\n");
exit(EXIT_WAIVED);
}
g_display = eglGetPlatformDisplayEXT(EGL_PLATFORM_DEVICE_EXT,
(void *)devices[egl_device_id], NULL);
if (g_display == EGL_NO_DISPLAY) {
printf("Could not get EGL display from device. \n");
eglStatus = EGL_FALSE;
exit(EXIT_FAILURE);
}
eglStatus = eglInitialize(g_display, 0, 0);
if (!eglStatus) {
printf("EGL failed to initialize. \n");
eglStatus = EGL_FALSE;
exit(EXIT_FAILURE);
}
eglStream = eglCreateStreamKHR(g_display, streamAttrMailboxMode);
if (eglStream == EGL_NO_STREAM_KHR) {
printf("Could not create EGL stream.\n");
eglStatus = EGL_FALSE;
exit(EXIT_FAILURE);
}
printf("Created EGLStream %p\n", eglStream);
// Set stream attribute
if (!eglStreamAttribKHR(g_display, eglStream, EGL_CONSUMER_LATENCY_USEC_KHR,
16000)) {
printf(
"Consumer: eglStreamAttribKHR EGL_CONSUMER_LATENCY_USEC_KHR failed\n");
return 0;
}
if (!eglStreamAttribKHR(g_display, eglStream,
EGL_CONSUMER_ACQUIRE_TIMEOUT_USEC_KHR, 16000)) {
printf(
"Consumer: eglStreamAttribKHR EGL_CONSUMER_ACQUIRE_TIMEOUT_USEC_KHR "
"failed\n");
return 0;
}
printf("EGLStream initialized\n");
return 1;
}
void EGLStreamFini(void) { eglDestroyStreamKHR(g_display, eglStream); }
#endif

View File

@@ -0,0 +1,103 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Common EGL stream functions header file
//
#ifndef _EGLSTRM_COMMON_H_
#define _EGLSTRM_COMMON_H_
#include <signal.h>
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <unistd.h>
#include "cuda.h"
#include "cudaEGL.h"
#include "helper_cuda_drvapi.h"
#define EXTENSION_LIST(T) \
T(PFNEGLCREATESTREAMKHRPROC, eglCreateStreamKHR) \
T(PFNEGLDESTROYSTREAMKHRPROC, eglDestroyStreamKHR) \
T(PFNEGLQUERYSTREAMKHRPROC, eglQueryStreamKHR) \
T(PFNEGLQUERYSTREAMU64KHRPROC, eglQueryStreamu64KHR) \
T(PFNEGLQUERYSTREAMTIMEKHRPROC, eglQueryStreamTimeKHR) \
T(PFNEGLSTREAMATTRIBKHRPROC, eglStreamAttribKHR) \
T(PFNEGLSTREAMCONSUMERACQUIREKHRPROC, eglStreamConsumerAcquireKHR) \
T(PFNEGLSTREAMCONSUMERRELEASEKHRPROC, eglStreamConsumerReleaseKHR) \
T(PFNEGLSTREAMCONSUMERGLTEXTUREEXTERNALKHRPROC, \
eglStreamConsumerGLTextureExternalKHR) \
T(PFNEGLGETSTREAMFILEDESCRIPTORKHRPROC, eglGetStreamFileDescriptorKHR) \
T(PFNEGLQUERYDEVICESEXTPROC, eglQueryDevicesEXT) \
T(PFNEGLGETPLATFORMDISPLAYEXTPROC, eglGetPlatformDisplayEXT) \
T(PFNEGLQUERYDEVICEATTRIBEXTPROC, eglQueryDeviceAttribEXT) \
T(PFNEGLCREATESTREAMFROMFILEDESCRIPTORKHRPROC, \
eglCreateStreamFromFileDescriptorKHR)
#define eglCreateStreamKHR my_eglCreateStreamKHR
#define eglDestroyStreamKHR my_eglDestroyStreamKHR
#define eglQueryStreamKHR my_eglQueryStreamKHR
#define eglQueryStreamu64KHR my_eglQueryStreamu64KHR
#define eglQueryStreamTimeKHR my_eglQueryStreamTimeKHR
#define eglStreamAttribKHR my_eglStreamAttribKHR
#define eglStreamConsumerAcquireKHR my_eglStreamConsumerAcquireKHR
#define eglStreamConsumerReleaseKHR my_eglStreamConsumerReleaseKHR
#define eglStreamConsumerGLTextureExternalKHR \
my_eglStreamConsumerGLTextureExternalKHR
#define eglGetStreamFileDescriptorKHR my_eglGetStreamFileDescriptorKHR
#define eglCreateStreamFromFileDescriptorKHR \
my_eglCreateStreamFromFileDescriptorKHR
#define eglQueryDevicesEXT my_eglQueryDevicesEXT
#define eglGetPlatformDisplayEXT my_eglGetPlatformDisplayEXT
#define eglQueryDeviceAttribEXT my_eglQueryDeviceAttribEXT
#define EXTLST_DECL(tx, x) tx my_##x = NULL;
#define EXTLST_EXTERN(tx, x) extern tx my_##x;
#define EXTLST_ENTRY(tx, x) {(extlst_fnptr_t *)&my_##x, #x},
#define MAX_STRING_SIZE 256
#define WIDTH 720
#define HEIGHT 480
typedef struct _TestArgs {
char *infile1;
char *infile2;
bool isARGB;
unsigned int inputWidth;
unsigned int inputHeight;
bool pitchLinearOutput;
} TestArgs;
int eglSetupExtensions(void);
int EGLStreamInit(int *dev);
void EGLStreamFini(void);
#endif

View File

@@ -0,0 +1,160 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# findegl.mk is used to find the necessary EGL Libraries for specific distributions
# this is supported on Linux
#
################################################################################
# Determine OS platform and unix distribution
ifeq ("$(TARGET_OS)","linux")
# first search lsb_release
DISTRO = $(shell lsb_release -i -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
ifeq ("$(DISTRO)","")
# second search and parse /etc/issue
DISTRO = $(shell more /etc/issue | awk '{print $$1}' | sed '1!d' | sed -e "/^$$/d" 2>/dev/null | tr "[:upper:]" "[:lower:]")
# ensure data from /etc/issue is valid
ifeq (,$(filter $(DISTRO),ubuntu fedora red rhel centos suse))
DISTRO =
endif
ifeq ("$(DISTRO)","")
# third, we can search in /etc/os-release or /etc/{distro}-release
DISTRO = $(shell awk '/ID/' /etc/*-release | sed 's/ID=//' | grep -v "VERSION" | grep -v "ID" | grep -v "DISTRIB")
endif
endif
endif
ifeq ("$(TARGET_OS)","linux")
# $(info) >> findegl.mk -> LINUX path <<<)
# Each set of Linux Distros have different paths for where to find their OpenGL libraries reside
UBUNTU = $(shell echo $(DISTRO) | grep -i ubuntu >/dev/null 2>&1; echo $$?)
FEDORA = $(shell echo $(DISTRO) | grep -i fedora >/dev/null 2>&1; echo $$?)
RHEL = $(shell echo $(DISTRO) | grep -i 'red\|rhel' >/dev/null 2>&1; echo $$?)
CENTOS = $(shell echo $(DISTRO) | grep -i centos >/dev/null 2>&1; echo $$?)
SUSE = $(shell echo $(DISTRO) | grep -i 'suse\|sles' >/dev/null 2>&1; echo $$?)
ifeq ("$(UBUNTU)","0")
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
GLPATH := /usr/arm-linux-gnueabihf/lib
GLLINK := -L/usr/arm-linux-gnueabihf/lib
ifneq ($(TARGET_FS),)
GLPATH += $(TARGET_FS)/usr/lib/arm-linux-gnueabihf
GLLINK += -L$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-aarch64)
GLPATH := /usr/aarch64-linux-gnu/lib
GLLINK := -L/usr/aarch64-linux-gnu/lib
ifneq ($(TARGET_FS),)
GLPATH += $(TARGET_FS)/usr/lib
GLPATH += $(TARGET_FS)/usr/lib/aarch64-linux-gnu
GLLINK += -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
endif
else
UBUNTU_PKG_NAME = $(shell which dpkg >/dev/null 2>&1 && dpkg -l 'nvidia-*' | grep '^ii' | awk '{print $$2}' | head -1)
ifneq ("$(UBUNTU_PKG_NAME)","")
GLPATH ?= /usr/lib/$(UBUNTU_PKG_NAME)
GLLINK ?= -L/usr/lib/$(UBUNTU_PKG_NAME)
endif
DFLT_PATH ?= /usr/lib
endif
endif
ifeq ("$(SUSE)","0")
GLPATH ?= /usr/X11R6/lib64
GLLINK ?= -L/usr/X11R6/lib64
DFLT_PATH ?= /usr/lib64
endif
ifeq ("$(FEDORA)","0")
GLPATH ?= /usr/lib64/nvidia
GLLINK ?= -L/usr/lib64/nvidia
DFLT_PATH ?= /usr/lib64
endif
ifeq ("$(RHEL)","0")
GLPATH ?= /usr/lib64/nvidia
GLLINK ?= -L/usr/lib64/nvidia
DFLT_PATH ?= /usr/lib64
endif
ifeq ("$(CENTOS)","0")
GLPATH ?= /usr/lib64/nvidia
GLLINK ?= -L/usr/lib64/nvidia
DFLT_PATH ?= /usr/lib64
endif
EGLLIB := $(shell find -L $(GLPATH) $(DFLT_PATH) -name libEGL.so -print 2>/dev/null)
ifeq ("$(EGLLIB)","")
$(info >>> WARNING - libEGL.so not found, please install libEGL.so <<<)
SAMPLE_ENABLED := 0
endif
HEADER_SEARCH_PATH ?= $(TARGET_FS)/usr/include
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
HEADER_SEARCH_PATH += /usr/arm-linux-gnueabihf/include
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-aarch64-linux)
HEADER_SEARCH_PATH += /usr/aarch64-linux-gnu/include
endif
EGLHEADER := $(shell find -L $(HEADER_SEARCH_PATH) -name egl.h -print 2>/dev/null)
EGLEXTHEADER := $(shell find -L $(HEADER_SEARCH_PATH) -name eglext.h -print 2>/dev/null)
ifeq ("$(EGLHEADER)","")
$(info >>> WARNING - egl.h not found, please install egl.h <<<)
SAMPLE_ENABLED := 0
endif
ifeq ("$(EGLEXTHEADER)","")
$(info >>> WARNING - eglext.h not found, please install eglext.h <<<)
SAMPLE_ENABLED := 0
endif
else
endif
ifeq ("$(TARGET_OS)","qnx")
HOST_CCFLAGS := -V5.4.0,gcc_ntoaarch64le
endif
# Attempt to compile a minimal EGL application and run to check if EGL_SUPPORT_REUSE_NV is supported in the EGL headers available.
ifneq ($(SAMPLE_ENABLED), 0)
$(shell printf "#include <EGL/egl.h>\n#include <EGL/eglext.h>\nint main() {\n#ifdef EGL_SUPPORT_REUSE_NV \n #error \"Compatible EGL header found\" \n return 0;\n#endif \n return 1;\n}" > test.c; )
EGL_DEFINES := $(shell $(HOST_COMPILER) $(HOST_CCFLAGS) $(CCFLAGS) $(EXTRA_CCFLAGS) -lEGL test.c -c 2>&1 | grep -ic "Compatible EGL header found";)
SHOULD_WAIVE := 0
ifeq ($(EGL_DEFINES),0)
SHOULD_WAIVE := 1
endif
ifeq ($(SHOULD_WAIVE),1)
$(info -----------------------------------------------------------------------------------------------)
$(info WARNING - NVIDIA EGL EXTENSIONS are not available in the present EGL headers)
$(info -----------------------------------------------------------------------------------------------)
$(info This CUDA Sample cannot be built if the EGL NVIDIA EXTENSIONS like EGL_SUPPORT_REUSE_NV are not supported in EGL headers.)
$(info This will be a dry-run of the Makefile.)
$(info Please install the latest khronos EGL headers and libs to build this sample)
$(info -----------------------------------------------------------------------------------------------)
SAMPLE_ENABLED := 0
endif
$(shell rm test.o test.c 2>/dev/null)
endif

View File

@@ -0,0 +1,231 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
//
// DESCRIPTION: Simple EGL stream sample app
//
//
//#define EGL_EGLEXT_PROTOTYPES
#include "cudaEGL.h"
#include "cuda_consumer.h"
#include "cuda_producer.h"
#include "eglstrm_common.h"
/* ------ globals ---------*/
#if defined(EXTENSION_LIST)
EXTENSION_LIST(EXTLST_EXTERN)
#endif
#define NUM_TRAILS 4
bool signal_stop = 0;
static void sig_handler(int sig) {
signal_stop = 1;
printf("Signal: %d\n", sig);
}
int main(int argc, char **argv) {
TestArgs args;
CUresult curesult = CUDA_SUCCESS;
unsigned int i, j;
EGLint streamState = 0;
test_cuda_consumer_s cudaConsumer;
test_cuda_producer_s cudaProducer;
memset(&cudaProducer, 0, sizeof(test_cuda_producer_s));
memset(&cudaConsumer, 0, sizeof(test_cuda_consumer_s));
// Hook up Ctrl-C handler
signal(SIGINT, sig_handler);
if (!eglSetupExtensions()) {
printf("SetupExtentions failed \n");
curesult = CUDA_ERROR_UNKNOWN;
goto done;
}
checkCudaErrors(cuInit(0));
int count;
checkCudaErrors(cuDeviceGetCount(&count));
printf("Found %d cuda devices\n", count);
CUdevice devId;
if (!EGLStreamInit(&devId)) {
printf("EGLStream Init failed.\n");
curesult = CUDA_ERROR_UNKNOWN;
goto done;
}
curesult = cudaDeviceCreateProducer(&cudaProducer, devId);
if (curesult != CUDA_SUCCESS) {
goto done;
}
curesult = cudaDeviceCreateConsumer(&cudaConsumer, devId);
if (curesult != CUDA_SUCCESS) {
goto done;
}
checkCudaErrors(cuCtxPushCurrent(cudaConsumer.context));
if (CUDA_SUCCESS != (curesult = cuEGLStreamConsumerConnect(
&(cudaConsumer.cudaConn), eglStream))) {
printf("FAILED Connect CUDA consumer with error %d\n", curesult);
goto done;
} else {
printf("Connected CUDA consumer, CudaConsumer %p\n", cudaConsumer.cudaConn);
}
checkCudaErrors(cuCtxPopCurrent(&cudaConsumer.context));
checkCudaErrors(cuCtxPushCurrent(cudaProducer.context));
if (CUDA_SUCCESS ==
(curesult = cuEGLStreamProducerConnect(&(cudaProducer.cudaConn),
eglStream, WIDTH, HEIGHT))) {
printf("Connect CUDA producer Done, CudaProducer %p\n",
cudaProducer.cudaConn);
} else {
printf("Connect CUDA producer FAILED with error %d\n", curesult);
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaProducer.context));
// Initialize producer
for (i = 0; i < NUM_TRAILS; i++) {
if (streamState != EGL_STREAM_STATE_CONNECTING_KHR) {
if (!eglQueryStreamKHR(g_display, eglStream, EGL_STREAM_STATE_KHR,
&streamState)) {
printf("main: eglQueryStreamKHR EGL_STREAM_STATE_KHR failed\n");
curesult = CUDA_ERROR_UNKNOWN;
goto done;
}
}
args.inputWidth = WIDTH;
args.inputHeight = HEIGHT;
if (i % 2 != 0) {
args.isARGB = 1;
args.infile1 = sdkFindFilePath("cuda_f_1.yuv", argv[0]);
args.infile2 = sdkFindFilePath("cuda_f_2.yuv", argv[0]);
} else {
args.isARGB = 0;
args.infile1 = sdkFindFilePath("cuda_yuv_f_1.yuv", argv[0]);
args.infile2 = sdkFindFilePath("cuda_yuv_f_2.yuv", argv[0]);
}
if ((i % 4) < 2) {
args.pitchLinearOutput = 1;
} else {
args.pitchLinearOutput = 0;
}
checkCudaErrors(cuCtxPushCurrent(cudaProducer.context));
cudaProducerInit(&cudaProducer, g_display, eglStream, &args);
checkCudaErrors(cuCtxPopCurrent(&cudaProducer.context));
checkCudaErrors(cuCtxPushCurrent(cudaConsumer.context));
cuda_consumer_init(&cudaConsumer, &args);
checkCudaErrors(cuCtxPopCurrent(&cudaConsumer.context));
printf("main - Cuda Producer and Consumer Initialized.\n");
for (j = 0; j < 2; j++) {
printf("Running for %s frame and %s input\n",
args.isARGB ? "ARGB" : "YUV",
args.pitchLinearOutput ? "Pitchlinear" : "BlockLinear");
if (j == 0) {
checkCudaErrors(cuCtxPushCurrent(cudaProducer.context));
curesult = cudaProducerTest(&cudaProducer, cudaProducer.fileName1);
if (curesult != CUDA_SUCCESS) {
printf("Cuda Producer Test failed for frame = %d\n", j + 1);
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaProducer.context));
checkCudaErrors(cuCtxPushCurrent(cudaConsumer.context));
curesult = cudaConsumerTest(&cudaConsumer, cudaConsumer.outFile1);
if (curesult != CUDA_SUCCESS) {
printf("Cuda Consumer Test failed for frame = %d\n", j + 1);
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaConsumer.context));
} else {
checkCudaErrors(cuCtxPushCurrent(cudaProducer.context));
curesult = cudaProducerTest(&cudaProducer, cudaProducer.fileName2);
if (curesult != CUDA_SUCCESS) {
printf("Cuda Producer Test failed for frame = %d\n", j + 1);
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaProducer.context));
checkCudaErrors(cuCtxPushCurrent(cudaConsumer.context));
curesult = cudaConsumerTest(&cudaConsumer, cudaConsumer.outFile2);
if (curesult != CUDA_SUCCESS) {
printf("Cuda Consumer Test failed for frame = %d\n", j + 1);
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaConsumer.context));
}
}
}
checkCudaErrors(cuCtxPushCurrent(cudaProducer.context));
if (CUDA_SUCCESS != (curesult = cudaProducerDeinit(&cudaProducer))) {
printf("Producer Disconnect FAILED. \n");
goto done;
}
checkCudaErrors(cuCtxPopCurrent(&cudaProducer.context));
if (!eglQueryStreamKHR(g_display, eglStream, EGL_STREAM_STATE_KHR,
&streamState)) {
printf("Cuda consumer, eglQueryStreamKHR EGL_STREAM_STATE_KHR failed\n");
curesult = CUDA_ERROR_UNKNOWN;
goto done;
}
if (streamState != EGL_STREAM_STATE_DISCONNECTED_KHR) {
if (CUDA_SUCCESS != (curesult = cuda_consumer_deinit(&cudaConsumer))) {
printf("Consumer Disconnect FAILED.\n");
goto done;
}
}
printf("Producer and Consumer Disconnected \n");
done:
if (!eglQueryStreamKHR(g_display, eglStream, EGL_STREAM_STATE_KHR,
&streamState)) {
printf("Cuda consumer, eglQueryStreamKHR EGL_STREAM_STATE_KHR failed\n");
curesult = CUDA_ERROR_UNKNOWN;
}
if (streamState != EGL_STREAM_STATE_DISCONNECTED_KHR) {
EGLStreamFini();
}
if (curesult == CUDA_SUCCESS) {
printf("&&&& EGLStream interop test PASSED\n");
} else {
printf("&&&& EGLStream interop test FAILED\n");
}
return 0;
}

View File

@@ -0,0 +1,341 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
SAMPLE_ENABLED := 1
# This sample is not supported on QNX
ifeq ($(TARGET_OS),qnx)
$(info >>> WARNING - MersenneTwisterGP11213 is not supported on QNX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# Gencode arguments
SMS ?=
ifeq ($(GENCODE_FLAGS),)
# Generate SASS code for each SM architecture listed in $(SMS)
$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
ifeq ($(SMS),)
# Generate PTX code from SM 35
GENCODE_FLAGS += -gencode arch=compute_35,code=compute_35
endif
# Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
HIGHEST_SM := $(lastword $(sort $(SMS)))
ifneq ($(HIGHEST_SM),)
GENCODE_FLAGS += -gencode arch=compute_$(HIGHEST_SM),code=compute_$(HIGHEST_SM)
endif
endif
ALL_CCFLAGS += --threads 0
LIBRARIES += -lcurand_static -lculibos
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: MersenneTwisterGP11213
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
MersenneTwister.o:MersenneTwister.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
MersenneTwisterGP11213: MersenneTwister.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./MersenneTwisterGP11213
clean:
rm -f MersenneTwisterGP11213 MersenneTwister.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/MersenneTwisterGP11213
clobber: clean

View File

@@ -0,0 +1,180 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/*
* This sample demonstrates the use of CURAND to generate
* random numbers on GPU and CPU.
*/
// Utilities and system includes
// includes, system
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <curand.h>
// Utilities and system includes
#include <helper_functions.h>
#include <helper_cuda.h>
#include <cuda_runtime.h>
#include <curand.h>
float compareResults(int rand_n, float *h_RandGPU, float *h_RandCPU);
const int DEFAULT_RAND_N = 2400000;
const unsigned int DEFAULT_SEED = 777;
///////////////////////////////////////////////////////////////////////////////
// Main program
///////////////////////////////////////////////////////////////////////////////
int main(int argc, char **argv) {
// Start logs
printf("%s Starting...\n\n", argv[0]);
// initialize the GPU, either identified by --device
// or by picking the device with highest flop rate.
int devID = findCudaDevice(argc, (const char **)argv);
// parsing the number of random numbers to generate
int rand_n = DEFAULT_RAND_N;
if (checkCmdLineFlag(argc, (const char **)argv, "count")) {
rand_n = getCmdLineArgumentInt(argc, (const char **)argv, "count");
}
printf("Allocating data for %i samples...\n", rand_n);
// parsing the seed
int seed = DEFAULT_SEED;
if (checkCmdLineFlag(argc, (const char **)argv, "seed")) {
seed = getCmdLineArgumentInt(argc, (const char **)argv, "seed");
}
printf("Seeding with %i ...\n", seed);
cudaStream_t stream;
checkCudaErrors(cudaStreamCreateWithFlags(&stream, cudaStreamNonBlocking));
float *d_Rand;
checkCudaErrors(cudaMalloc((void **)&d_Rand, rand_n * sizeof(float)));
curandGenerator_t prngGPU;
checkCudaErrors(curandCreateGenerator(&prngGPU, CURAND_RNG_PSEUDO_MTGP32));
checkCudaErrors(curandSetStream(prngGPU, stream));
checkCudaErrors(curandSetPseudoRandomGeneratorSeed(prngGPU, seed));
curandGenerator_t prngCPU;
checkCudaErrors(
curandCreateGeneratorHost(&prngCPU, CURAND_RNG_PSEUDO_MTGP32));
checkCudaErrors(curandSetPseudoRandomGeneratorSeed(prngCPU, seed));
//
// Example 1: Compare random numbers generated on GPU and CPU
float *h_RandGPU;
checkCudaErrors(cudaMallocHost(&h_RandGPU, rand_n * sizeof(float)));
printf("Generating random numbers on GPU...\n\n");
checkCudaErrors(curandGenerateUniform(prngGPU, (float *)d_Rand, rand_n));
printf("\nReading back the results...\n");
checkCudaErrors(cudaMemcpyAsync(h_RandGPU, d_Rand, rand_n * sizeof(float),
cudaMemcpyDeviceToHost, stream));
float *h_RandCPU = (float *)malloc(rand_n * sizeof(float));
printf("Generating random numbers on CPU...\n\n");
checkCudaErrors(curandGenerateUniform(prngCPU, (float *)h_RandCPU, rand_n));
checkCudaErrors(cudaStreamSynchronize(stream));
printf("Comparing CPU/GPU random numbers...\n\n");
float L1norm = compareResults(rand_n, h_RandGPU, h_RandCPU);
//
// Example 2: Timing of random number generation on GPU
const int numIterations = 10;
int i;
StopWatchInterface *hTimer;
sdkCreateTimer(&hTimer);
sdkResetTimer(&hTimer);
sdkStartTimer(&hTimer);
for (i = 0; i < numIterations; i++) {
checkCudaErrors(curandGenerateUniform(prngGPU, (float *)d_Rand, rand_n));
}
checkCudaErrors(cudaStreamSynchronize(stream));
sdkStopTimer(&hTimer);
double gpuTime = 1.0e-3 * sdkGetTimerValue(&hTimer) / (double)numIterations;
printf(
"MersenneTwisterGP11213, Throughput = %.4f GNumbers/s, Time = %.5f s, "
"Size = %u Numbers\n",
1.0e-9 * rand_n / gpuTime, gpuTime, rand_n);
printf("Shutting down...\n");
checkCudaErrors(curandDestroyGenerator(prngGPU));
checkCudaErrors(curandDestroyGenerator(prngCPU));
checkCudaErrors(cudaStreamDestroy(stream));
checkCudaErrors(cudaFree(d_Rand));
sdkDeleteTimer(&hTimer);
checkCudaErrors(cudaFreeHost(h_RandGPU));
free(h_RandCPU);
exit(L1norm < 1e-6 ? EXIT_SUCCESS : EXIT_FAILURE);
}
float compareResults(int rand_n, float *h_RandGPU, float *h_RandCPU) {
int i;
float rCPU, rGPU, delta;
float max_delta = 0.;
float sum_delta = 0.;
float sum_ref = 0.;
for (i = 0; i < rand_n; i++) {
rCPU = h_RandCPU[i];
rGPU = h_RandGPU[i];
delta = fabs(rCPU - rGPU);
sum_delta += delta;
sum_ref += fabs(rCPU);
if (delta >= max_delta) {
max_delta = delta;
}
}
float L1norm = (float)(sum_delta / sum_ref);
printf("Max absolute error: %E\n", max_delta);
printf("L1 norm: %E\n\n", L1norm);
return L1norm;
}

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "MersenneTwisterGP11213", "MersenneTwisterGP11213_vs2017.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,113 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>MersenneTwisterGP11213_vs2017</RootNamespace>
<ProjectName>MersenneTwisterGP11213</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<PropertyGroup Condition="'$(WindowsTargetPlatformVersion)'==''">
<LatestTargetPlatformVersion>$([Microsoft.Build.Utilities.ToolLocationHelper]::GetLatestSDKTargetPlatformVersion('Windows', '10.0'))</LatestTargetPlatformVersion>
<WindowsTargetPlatformVersion Condition="'$(WindowsTargetPlatformVersion)' == ''">$(LatestTargetPlatformVersion)</WindowsTargetPlatformVersion>
<TargetPlatformVersion>$(WindowsTargetPlatformVersion)</TargetPlatformVersion>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v141</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;$(CudaToolkitIncludeDir);</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>curand.lib;cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/MersenneTwisterGP11213.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,compute_35;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="MersenneTwister.cpp" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2019
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "MersenneTwisterGP11213", "MersenneTwisterGP11213_vs2019.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,109 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>MersenneTwisterGP11213_vs2019</RootNamespace>
<ProjectName>MersenneTwisterGP11213</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v142</PlatformToolset>
<WindowsTargetPlatformVersion>10.0</WindowsTargetPlatformVersion>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;$(CudaToolkitIncludeDir);</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>curand.lib;cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/MersenneTwisterGP11213.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,compute_35;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="MersenneTwister.cpp" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,74 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>MersenneTwisterGP11213</name>
<description><![CDATA[This sample demonstrates the Mersenne Twister random number generator GP11213 in cuRAND.]]></description>
<devicecompilation>whole</devicecompilation>
<fallback_min_ptx>true</fallback_min_ptx>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../common/inc</path>
</includepaths>
<keyconcepts>
<concept level="basic">CURAND Library</concept>
</keyconcepts>
<keywords>
<keyword>CUDA</keyword>
<keyword>CURAND</keyword>
<keyword>Monte-Carlo</keyword>
<keyword>random number generation</keyword>
<keyword>GSFR</keyword>
<keyword>Mersenne Twister</keyword>
</keywords>
<libraries>
<library>curand_static</library>
<library>culibos</library>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>MersenneTwister.cpp</primary_file>
<required_dependencies>
<dependency>CURAND</dependency>
</required_dependencies>
<scopes>
<scope>1:CUDA Advanced Topics</scope>
</scopes>
<sm-arch>sm35</sm-arch>
<sm-arch>sm37</sm-arch>
<sm-arch>sm50</sm-arch>
<sm-arch>sm52</sm-arch>
<sm-arch>sm60</sm-arch>
<sm-arch>sm61</sm-arch>
<sm-arch>sm70</sm-arch>
<sm-arch>sm72</sm-arch>
<sm-arch>sm75</sm-arch>
<sm-arch>sm80</sm-arch>
<sm-arch>sm86</sm-arch>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<platform>windows7</platform>
</env>
<env>
<arch>x86_64</arch>
<platform>macosx</platform>
</env>
<env>
<arch>arm</arch>
</env>
<env>
<arch>ppc64le</arch>
<platform>linux</platform>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>MersenneTwisterGP11213</title>
<type>exe</type>
</entry>

View File

@@ -0,0 +1,71 @@
# MersenneTwisterGP11213 - MersenneTwisterGP11213
## Description
This sample demonstrates the Mersenne Twister random number generator GP11213 in cuRAND.
## Key Concepts
CURAND Library
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux, Windows
## Supported CPU Architecture
x86_64, ppc64le, armv7l
## CUDA APIs involved
## Dependencies needed to build/run
[CURAND](../../README.md#curand)
## Prerequisites
Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run
### Windows
The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
```
*_vs<version>.sln - for Visual Studio <version>
```
Each individual sample has its own set of solution files in its directory:
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check DirectX Dependencies section for details."
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, armv7l.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=armv7l` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,355 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
SAMPLE_ENABLED := 1
# This sample is not supported on ARMv7
ifeq ($(TARGET_ARCH),armv7l)
$(info >>> WARNING - NV12toBGRandResize is not supported on ARMv7 - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# Gencode arguments
ifeq ($(TARGET_ARCH),$(filter $(TARGET_ARCH),armv7l aarch64))
SMS ?= 35 37 50 52 60 61 70 72 75 80 86
else
SMS ?= 35 37 50 52 60 61 70 75 80 86
endif
ifeq ($(SMS),)
$(info >>> WARNING - no SM architectures have been specified - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ifeq ($(GENCODE_FLAGS),)
# Generate SASS code for each SM architecture listed in $(SMS)
$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
# Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
HIGHEST_SM := $(lastword $(sort $(SMS)))
ifneq ($(HIGHEST_SM),)
GENCODE_FLAGS += -gencode arch=compute_$(HIGHEST_SM),code=compute_$(HIGHEST_SM)
endif
endif
ALL_CCFLAGS += --threads 0
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: NV12toBGRandResize
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
bgr_resize.o:bgr_resize.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
nv12_resize.o:nv12_resize.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
nv12_to_bgr_planar.o:nv12_to_bgr_planar.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
resize_convert_main.o:resize_convert_main.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
utils.o:utils.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
NV12toBGRandResize: bgr_resize.o nv12_resize.o nv12_to_bgr_planar.o resize_convert_main.o utils.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./NV12toBGRandResize
clean:
rm -f NV12toBGRandResize bgr_resize.o nv12_resize.o nv12_to_bgr_planar.o resize_convert_main.o utils.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/NV12toBGRandResize
clobber: clean

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "NV12toBGRandResize", "NV12toBGRandResize_vs2017.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,118 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>NV12toBGRandResize_vs2017</RootNamespace>
<ProjectName>NV12toBGRandResize</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<PropertyGroup Condition="'$(WindowsTargetPlatformVersion)'==''">
<LatestTargetPlatformVersion>$([Microsoft.Build.Utilities.ToolLocationHelper]::GetLatestSDKTargetPlatformVersion('Windows', '10.0'))</LatestTargetPlatformVersion>
<WindowsTargetPlatformVersion Condition="'$(WindowsTargetPlatformVersion)' == ''">$(LatestTargetPlatformVersion)</WindowsTargetPlatformVersion>
<TargetPlatformVersion>$(WindowsTargetPlatformVersion)</TargetPlatformVersion>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v141</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/NV12toBGRandResize.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,sm_35;compute_37,sm_37;compute_50,sm_50;compute_52,sm_52;compute_60,sm_60;compute_61,sm_61;compute_70,sm_70;compute_75,sm_75;compute_80,sm_80;compute_86,sm_86;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<CudaCompile Include="bgr_resize.cu" />
<CudaCompile Include="nv12_resize.cu" />
<CudaCompile Include="nv12_to_bgr_planar.cu" />
<ClCompile Include="resize_convert_main.cpp" />
<CudaCompile Include="utils.cu" />
<ClInclude Include="resize_convert.h" />
<ClInclude Include="utils.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2019
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "NV12toBGRandResize", "NV12toBGRandResize_vs2019.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,114 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>NV12toBGRandResize_vs2019</RootNamespace>
<ProjectName>NV12toBGRandResize</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v142</PlatformToolset>
<WindowsTargetPlatformVersion>10.0</WindowsTargetPlatformVersion>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/NV12toBGRandResize.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,sm_35;compute_37,sm_37;compute_50,sm_50;compute_52,sm_52;compute_60,sm_60;compute_61,sm_61;compute_70,sm_70;compute_75,sm_75;compute_80,sm_80;compute_86,sm_86;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<CudaCompile Include="bgr_resize.cu" />
<CudaCompile Include="nv12_resize.cu" />
<CudaCompile Include="nv12_to_bgr_planar.cu" />
<ClCompile Include="resize_convert_main.cpp" />
<CudaCompile Include="utils.cu" />
<ClInclude Include="resize_convert.h" />
<ClInclude Include="utils.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

View File

@@ -0,0 +1,71 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>NV12toBGRandResize</name>
<cuda_api_list>
<toolkit>cudaMemcpy2D</toolkit>
<toolkit>cudaMallocManaged</toolkit>
</cuda_api_list>
<description><![CDATA[This code shows two ways to convert and resize NV12 frames to BGR 3 planars frames using CUDA in batch. Way-1, Convert NV12 Input to BGR @ Input Resolution-1, then Resize to Resolution#2. Way-2, resize NV12 Input to Resolution#2 then convert it to BGR Output. NVIDIA HW Decoder, both dGPU and Tegra, normally outputs NV12 pitch format frames. For the inference using TensorRT, the input frame needs to be BGR planar format with possibly different size. So, conversion and resizing from NV12 to BGR planar is usually required for the inference following decoding. This CUDA code provides a reference implementation for conversion and resizing.]]></description>
<devicecompilation>whole</devicecompilation>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../common/inc</path>
</includepaths>
<keyconcepts>
<concept level="basic">Graphics Interop</concept>
<concept level="basic">Image Processing</concept>
<concept level="basic">Video Processing</concept>
</keyconcepts>
<keywords>
<keyword>GPGPU</keyword>
</keywords>
<libraries>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>resize_convert_main.cpp</primary_file>
<scopes>
<scope>1:CUDA Basic Topics</scope>
<scope>2:Image Processing</scope>
<scope>2:Computer Vision</scope>
</scopes>
<sm-arch>sm35</sm-arch>
<sm-arch>sm37</sm-arch>
<sm-arch>sm50</sm-arch>
<sm-arch>sm52</sm-arch>
<sm-arch>sm60</sm-arch>
<sm-arch>sm61</sm-arch>
<sm-arch>sm70</sm-arch>
<sm-arch>sm72</sm-arch>
<sm-arch>sm75</sm-arch>
<sm-arch>sm80</sm-arch>
<sm-arch>sm86</sm-arch>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<platform>windows7</platform>
</env>
<env>
<arch>x86_64</arch>
<platform>macosx</platform>
</env>
<env>
<arch>aarch64</arch>
</env>
<env>
<arch>ppc64le</arch>
<platform>linux</platform>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>NV12toBGRandResize</title>
<type>exe</type>
</entry>

View File

@@ -0,0 +1,70 @@
# NV12toBGRandResize - NV12toBGRandResize
## Description
This code shows two ways to convert and resize NV12 frames to BGR 3 planars frames using CUDA in batch. Way-1, Convert NV12 Input to BGR @ Input Resolution-1, then Resize to Resolution#2. Way-2, resize NV12 Input to Resolution#2 then convert it to BGR Output. NVIDIA HW Decoder, both dGPU and Tegra, normally outputs NV12 pitch format frames. For the inference using TensorRT, the input frame needs to be BGR planar format with possibly different size. So, conversion and resizing from NV12 to BGR planar is usually required for the inference following decoding. This CUDA code provides a reference implementation for conversion and resizing.
## Key Concepts
Graphics Interop, Image Processing, Video Processing
## Supported SM Architectures
[SM 3.5 ](https://developer.nvidia.com/cuda-gpus) [SM 3.7 ](https://developer.nvidia.com/cuda-gpus) [SM 5.0 ](https://developer.nvidia.com/cuda-gpus) [SM 5.2 ](https://developer.nvidia.com/cuda-gpus) [SM 6.0 ](https://developer.nvidia.com/cuda-gpus) [SM 6.1 ](https://developer.nvidia.com/cuda-gpus) [SM 7.0 ](https://developer.nvidia.com/cuda-gpus) [SM 7.2 ](https://developer.nvidia.com/cuda-gpus) [SM 7.5 ](https://developer.nvidia.com/cuda-gpus) [SM 8.0 ](https://developer.nvidia.com/cuda-gpus) [SM 8.6 ](https://developer.nvidia.com/cuda-gpus)
## Supported OSes
Linux, Windows
## Supported CPU Architecture
x86_64, ppc64le, aarch64
## CUDA APIs involved
### [CUDA Runtime API](http://docs.nvidia.com/cuda/cuda-runtime-api/index.html)
cudaMemcpy2D, cudaMallocManaged
## Prerequisites
Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
## Build and Run
### Windows
The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
```
*_vs<version>.sln - for Visual Studio <version>
```
Each individual sample has its own set of solution files in its directory:
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check DirectX Dependencies section for details."
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, aarch64.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=aarch64` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,134 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
// Implements BGR 3 progressive planars frames batch resize
#include <cuda.h>
#include <cuda_runtime.h>
#include "resize_convert.h"
__global__ void resizeBGRplanarBatchKernel(cudaTextureObject_t texSrc,
float *pDst, int nDstPitch, int nDstHeight, int nSrcHeight,
int batch, float scaleX, float scaleY,
int cropX, int cropY, int cropW, int cropH) {
int x = threadIdx.x + blockIdx.x * blockDim.x;
int y = threadIdx.y + blockIdx.y * blockDim.y;
if (x >= (int)(cropW/scaleX) || y >= (int)(cropH/scaleY))
return;
int frameSize = nDstPitch*nDstHeight;
float *p = NULL;
for (int i = blockIdx.z; i < batch; i += gridDim.z) {
#pragma unroll
for (int channel=0; channel < 3; channel++){
p = pDst + i * 3 * frameSize + y * nDstPitch + x + channel * frameSize;
*p = tex2D<float>(texSrc, x * scaleX + cropX,
((3 * i + channel) * nSrcHeight + y * scaleY + cropY));
}
}
}
static void resizeBGRplanarBatchCore(
float *dpSrc, int nSrcPitch, int nSrcWidth, int nSrcHeight,
float *dpDst, int nDstPitch, int nDstWidth, int nDstHeight,
int nBatchSize, cudaStream_t stream, bool whSameResizeRatio,
int cropX, int cropY, int cropW, int cropH) {
cudaTextureObject_t texSrc[2];
int nTiles = 1, h, iTile;
h = nSrcHeight * 3 * nBatchSize;
while ((h + nTiles - 1) / nTiles > 65536)
nTiles++;
if (nTiles > 2)
return;
int batchTile = nBatchSize / nTiles;
int batchTileLast = nBatchSize - batchTile * (nTiles-1);
for (iTile = 0; iTile < nTiles; ++iTile) {
int bs = (iTile == nTiles - 1) ? batchTileLast : batchTile;
float *dpSrcNew = dpSrc +
iTile * (batchTile * 3 * nSrcHeight * nSrcPitch);
cudaResourceDesc resDesc = {};
resDesc.resType = cudaResourceTypePitch2D;
resDesc.res.pitch2D.devPtr = dpSrcNew;
resDesc.res.pitch2D.desc = cudaCreateChannelDesc<float>();
resDesc.res.pitch2D.width = nSrcWidth;
resDesc.res.pitch2D.height = bs * 3 * nSrcHeight;
resDesc.res.pitch2D.pitchInBytes = nSrcPitch * sizeof(float);
cudaTextureDesc texDesc = {};
texDesc.filterMode = cudaFilterModeLinear;
texDesc.readMode = cudaReadModeElementType;
checkCudaErrors(cudaCreateTextureObject(&texSrc[iTile], &resDesc, &texDesc, NULL));
float *dpDstNew = dpDst +
iTile * (batchTile * 3 * nDstHeight * nDstPitch);
if(cropW == 0 || cropH == 0) {
cropX = 0;
cropY = 0;
cropW = nSrcWidth;
cropH = nSrcHeight;
}
float scaleX = (cropW*1.0f / nDstWidth);
float scaleY = (cropH*1.0f / nDstHeight);
if(whSameResizeRatio == true)
scaleX = scaleY = scaleX > scaleY ? scaleX : scaleY;
dim3 block(32, 32, 1);
size_t blockDimZ = bs;
// Restricting blocks in Z-dim till 32 to not launch too many blocks
blockDimZ = (blockDimZ > 32) ? 32 : blockDimZ;
dim3 grid((cropW*1.0f/scaleX + block.x - 1) / block.x,
(cropH*1.0f/scaleY + block.y - 1) / block.y, blockDimZ);
resizeBGRplanarBatchKernel<<<grid, block, 0, stream>>>
(texSrc[iTile], dpDstNew, nDstPitch, nDstHeight, nSrcHeight,
bs, scaleX, scaleY, cropX, cropY, cropW, cropH);
}
for (iTile = 0; iTile < nTiles; ++iTile)
checkCudaErrors(cudaDestroyTextureObject(texSrc[iTile]));
}
void resizeBGRplanarBatch(
float *dpSrc, int nSrcPitch, int nSrcWidth, int nSrcHeight,
float *dpDst, int nDstPitch, int nDstWidth, int nDstHeight,
int nBatchSize, cudaStream_t stream,
int cropX, int cropY, int cropW, int cropH, bool whSameResizeRatio) {
resizeBGRplanarBatchCore(dpSrc, nSrcPitch, nSrcWidth, nSrcHeight,
dpDst, nDstPitch, nDstWidth, nDstHeight, nBatchSize, stream,
whSameResizeRatio, cropX, cropY, cropW, cropH);
}

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

View File

@@ -0,0 +1,112 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
// Implements interlace NV12 frames batch resize
#include <cuda.h>
#include <cuda_runtime.h>
#include "resize_convert.h"
__global__ static void resizeNV12BatchKernel(cudaTextureObject_t texSrcLuma,
cudaTextureObject_t texSrcChroma,
uint8_t *pDstNv12, int nSrcWidth,
int nSrcHeight, int nDstPitch,
int nDstWidth, int nDstHeight,
int nBatchSize) {
int x = threadIdx.x + blockIdx.x * blockDim.x;
int y = threadIdx.y + blockIdx.y * blockDim.y;
int px = x * 2, py = y * 2;
if ((px + 1) >= nDstWidth || (py + 1) >= nDstHeight) return;
float fxScale = 1.0f * nSrcWidth / nDstWidth;
float fyScale = 1.0f * nSrcHeight / nDstHeight;
uint8_t *p = pDstNv12 + px + py * nDstPitch;
int hh = nDstHeight * 3 / 2;
int nByte = nDstPitch * hh;
int px_fxScale = px * fxScale;
int px_fxScale_1 = (px + 1) * fxScale;
int py_fyScale = py * fyScale;
int py_fyScale_1 = (py + 1) * fyScale;
for (int i = blockIdx.z; i < nBatchSize; i+=gridDim.z) {
*(uchar2 *)p = make_uchar2(tex2D<uint8_t>(texSrcLuma, px_fxScale, py_fyScale),
tex2D<uint8_t>(texSrcLuma, px_fxScale_1, py_fyScale));
*(uchar2 *)(p + nDstPitch) =
make_uchar2(tex2D<uint8_t>(texSrcLuma, px_fxScale, py_fyScale_1),
tex2D<uint8_t>(texSrcLuma, px_fxScale_1, py_fyScale_1));
*(uchar2 *)(p + (nDstHeight - y) * nDstPitch) = tex2D<uchar2>(
texSrcChroma, x * fxScale, (hh * i + nDstHeight + y) * fyScale);
p += nByte;
py += hh;
}
}
void resizeNV12Batch(uint8_t *dpSrc, int nSrcPitch, int nSrcWidth,
int nSrcHeight, uint8_t *dpDst, int nDstPitch,
int nDstWidth, int nDstHeight, int nBatchSize,
cudaStream_t stream) {
int hhSrc = ceilf(nSrcHeight * 3.0f / 2.0f);
cudaResourceDesc resDesc = {};
resDesc.resType = cudaResourceTypePitch2D;
resDesc.res.pitch2D.devPtr = dpSrc;
resDesc.res.pitch2D.desc = cudaCreateChannelDesc<uint8_t>();
resDesc.res.pitch2D.width = nSrcWidth;
resDesc.res.pitch2D.height = hhSrc * nBatchSize;
resDesc.res.pitch2D.pitchInBytes = nSrcPitch;
cudaTextureDesc texDesc = {};
texDesc.filterMode = cudaFilterModePoint;
texDesc.readMode = cudaReadModeElementType;
cudaTextureObject_t texLuma = 0;
checkCudaErrors(cudaCreateTextureObject(&texLuma, &resDesc, &texDesc, NULL));
resDesc.res.pitch2D.desc = cudaCreateChannelDesc<uchar2>();
resDesc.res.pitch2D.width /= 2;
cudaTextureObject_t texChroma = 0;
checkCudaErrors(cudaCreateTextureObject(&texChroma, &resDesc, &texDesc, NULL));
dim3 block(32, 32, 1);
size_t blockDimZ = nBatchSize;
// Restricting blocks in Z-dim till 32 to not launch too many blocks
blockDimZ = (blockDimZ > 32) ? 32 : blockDimZ;
dim3 grid((nDstWidth / 2 + block.x) / block.x,
(nDstHeight / 2 + block.y) / block.y, blockDimZ);
resizeNV12BatchKernel<<<grid, block, 0, stream>>>(
texLuma, texChroma, dpDst, nSrcWidth, nSrcHeight, nDstPitch, nDstWidth,
nDstHeight, nBatchSize);
checkCudaErrors(cudaDestroyTextureObject(texLuma));
checkCudaErrors(cudaDestroyTextureObject(texChroma));
}

View File

@@ -0,0 +1,154 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
// Implements NV12 to BGR batch conversion
#include <cuda.h>
#include <cuda_runtime.h>
#include "resize_convert.h"
#define CONV_THREADS_X 64
#define CONV_THREADS_Y 10
__forceinline__ __device__ static float clampF(float x, float lower,
float upper) {
return x < lower ? lower : (x > upper ? upper : x);
}
__global__ static void nv12ToBGRplanarBatchKernel(const uint8_t *pNv12,
int nNv12Pitch, float *pBgr,
int nRgbPitch, int nWidth,
int nHeight, int nBatchSize) {
int x = threadIdx.x + blockIdx.x * blockDim.x;
int y = threadIdx.y + blockIdx.y * blockDim.y;
if ((x << 2) + 1 > nWidth || (y << 1) + 1 > nHeight) return;
const uint8_t *__restrict__ pSrc = pNv12;
for (int i = blockIdx.z; i < nBatchSize; i += gridDim.z) {
pSrc = pNv12 + i * ((nHeight * nNv12Pitch * 3) >> 1) + (x << 2) +
(y << 1) * nNv12Pitch;
uchar4 luma2x01, luma2x23, uv2;
*(uint32_t *)&luma2x01 = *(uint32_t *)pSrc;
*(uint32_t *)&luma2x23 = *(uint32_t *)(pSrc + nNv12Pitch);
*(uint32_t *)&uv2 = *(uint32_t *)(pSrc + (nHeight - y) * nNv12Pitch);
float *pDstBlock = (pBgr + i * ((nHeight * nRgbPitch * 3) >> 2) +
((blockIdx.x * blockDim.x) << 2) +
((blockIdx.y * blockDim.y) << 1) * (nRgbPitch >> 2));
float2 add1;
float2 add2;
float2 add3;
float2 add00, add01, add02, add03;
float2 d, e;
add00.x = 1.1644f * luma2x01.x;
add01.x = 1.1644f * luma2x01.y;
add00.y = 1.1644f * luma2x01.z;
add01.y = 1.1644f * luma2x01.w;
add02.x = 1.1644f * luma2x23.x;
add03.x = 1.1644f * luma2x23.y;
add02.y = 1.1644f * luma2x23.z;
add03.y = 1.1644f * luma2x23.w;
d.x = uv2.x - 128.0f;
e.x = uv2.y - 128.0f;
d.y = uv2.z - 128.0f;
e.y = uv2.w - 128.0f;
add1.x = 2.0172f * d.x;
add1.y = 2.0172f * d.y;
add2.x = (-0.3918f) * d.x + (-0.8130f) * e.x;
add2.y = (-0.3918f) * d.y + (-0.8130f) * e.y;
add3.x = 1.5960f * e.x;
add3.y = 1.5960f * e.y;
int rowStride = (threadIdx.y << 1) * (nRgbPitch >> 2);
int nextRowStride = ((threadIdx.y << 1) + 1) * (nRgbPitch >> 2);
// B
*((float4 *)&pDstBlock[rowStride + (threadIdx.x << 2)]) =
make_float4(clampF(add00.x + add1.x, 0.0f, 255.0f),
clampF(add01.x + add1.x, 0.0f, 255.0f),
clampF(add00.y + add1.y, 0.0f, 255.0f),
clampF(add01.y + add1.y, 0.0f, 255.0f));
*((float4 *)&pDstBlock[nextRowStride + (threadIdx.x << 2)]) =
make_float4(clampF(add02.x + add1.x, 0.0f, 255.0f),
clampF(add03.x + add1.x, 0.0f, 255.0f),
clampF(add02.y + add1.y, 0.0f, 255.0f),
clampF(add03.y + add1.y, 0.0f, 255.0f));
int planeStride = nHeight * nRgbPitch >> 2;
// G
*((float4 *)&pDstBlock[planeStride + rowStride + (threadIdx.x << 2)]) =
make_float4(clampF(add00.x + add2.x, 0.0f, 255.0f),
clampF(add01.x + add2.x, 0.0f, 255.0f),
clampF(add00.y + add2.y, 0.0f, 255.0f),
clampF(add01.y + add2.y, 0.0f, 255.0f));
*((float4 *)&pDstBlock[planeStride + nextRowStride + (threadIdx.x << 2)]) =
make_float4(clampF(add02.x + add2.x, 0.0f, 255.0f),
clampF(add03.x + add2.x, 0.0f, 255.0f),
clampF(add02.y + add2.y, 0.0f, 255.0f),
clampF(add03.y + add2.y, 0.0f, 255.0f));
// R
*((float4
*)&pDstBlock[(planeStride << 1) + rowStride + (threadIdx.x << 2)]) =
make_float4(clampF(add00.x + add3.x, 0.0f, 255.0f),
clampF(add01.x + add3.x, 0.0f, 255.0f),
clampF(add00.y + add3.y, 0.0f, 255.0f),
clampF(add01.y + add3.y, 0.0f, 255.0f));
*((float4 *)&pDstBlock[(planeStride << 1) + nextRowStride +
(threadIdx.x << 2)]) =
make_float4(clampF(add02.x + add3.x, 0.0f, 255.0f),
clampF(add03.x + add3.x, 0.0f, 255.0f),
clampF(add02.y + add3.y, 0.0f, 255.0f),
clampF(add03.y + add3.y, 0.0f, 255.0f));
}
}
void nv12ToBGRplanarBatch(uint8_t *pNv12, int nNv12Pitch, float *pBgr,
int nRgbPitch, int nWidth, int nHeight,
int nBatchSize, cudaStream_t stream) {
dim3 threads(CONV_THREADS_X, CONV_THREADS_Y);
size_t blockDimZ = nBatchSize;
// Restricting blocks in Z-dim till 32 to not launch too many blocks
blockDimZ = (blockDimZ > 32) ? 32 : blockDimZ;
dim3 blocks((nWidth / 4 - 1) / threads.x + 1,
(nHeight / 2 - 1) / threads.y + 1, blockDimZ);
nv12ToBGRplanarBatchKernel<<<blocks, threads, 0, stream>>>(
pNv12, nNv12Pitch, pBgr, nRgbPitch, nWidth, nHeight, nBatchSize);
}

View File

@@ -0,0 +1,56 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __H_RESIZE_CONVERT__
#define __H_RESIZE_CONVERT__
#include <iostream>
#include <helper_cuda.h>
// nv12 resize
extern "C"
void resizeNV12Batch(
uint8_t *dpSrc, int nSrcPitch, int nSrcWidth, int nSrcHeight,
uint8_t *dpDst, int nDstPitch, int nDstWidth, int nDstHeight,
int nBatchSize, cudaStream_t stream = 0);
// bgr resize
extern "C"
void resizeBGRplanarBatch(
float *dpSrc, int nSrcPitch, int nSrcWidth, int nSrcHeight,
float *dpDst, int nDstPitch, int nDstWidth, int nDstHeight,
int nBatchSize, cudaStream_t stream = 0,
int cropX = 0, int cropY = 0, int cropW = 0, int cropH = 0,
bool whSameResizeRatio = false);
//NV12 to bgr planar
extern "C"
void nv12ToBGRplanarBatch(uint8_t *pNv12, int nNv12Pitch,
float *pRgb, int nRgbPitch, int nWidth, int nHeight,
int nBatchSize, cudaStream_t stream=0);
#endif

View File

@@ -0,0 +1,448 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/*
NVIDIA HW Decoder, both dGPU and Tegra, normally outputs NV12 pitch format
frames. For the inference using TensorRT, the input frame needs to be BGR planar
format with possibly different size. So, conversion and resizing from NV12 to
BGR planar is usually required for the inference following decoding.
This CUDA code is to provide a reference implementation for conversion and
resizing.
Limitaion
=========
NV12resize needs the height to be a even value.
Note
====
Resize function needs the pitch of image buffer to be 32 alignment.
Run
====
./NV12toBGRandResize
OR
./NV12toBGRandResize -input=data/test1920x1080.nv12 -width=1920 -height=1080 \
-dst_width=640 -dst_height=480 -batch=40 -device=0
*/
#include <cuda.h>
#include <cuda_runtime.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <cassert>
#include <fstream>
#include <iostream>
#include <memory>
#include "resize_convert.h"
#include "utils.h"
#define TEST_LOOP 20
typedef struct _nv12_to_bgr24_context_t {
int width;
int height;
int pitch;
int dst_width;
int dst_height;
int dst_pitch;
int batch;
int device; // cuda device ID
char *input_nv12_file;
int ctx_pitch; // the value will be suitable for Texture memroy.
int ctx_heights; // the value will be even.
} nv12_to_bgr24_context;
nv12_to_bgr24_context g_ctx;
static void printHelp(const char *app_name) {
std::cout << "Usage:" << app_name << " [options]\n\n";
std::cout << "OPTIONS:\n";
std::cout << "\t-h,--help\n\n";
std::cout << "\t-input=nv12file nv12 input file\n";
std::cout
<< "\t-width=width input nv12 image width, <1 -- 4096>\n";
std::cout
<< "\t-height=height input nv12 image height, <1 -- 4096>\n";
std::cout
<< "\t-pitch=pitch(optional) input nv12 image pitch, <0 -- 4096>\n";
std::cout
<< "\t-dst_width=width output BGR image width, <1 -- 4096>\n";
std::cout
<< "\t-dst_height=height output BGR image height, <1 -- 4096>\n";
std::cout
<< "\t-dst_pitch=pitch(optional) output BGR image pitch, <0 -- 4096>\n";
std::cout
<< "\t-batch=batch process frames count, <1 -- 4096>\n\n";
std::cout
<< "\t-device=device_num(optional) cuda device number, <0 -- 4096>\n\n";
return;
}
int parseCmdLine(int argc, char *argv[]) {
char **argp = (char **)argv;
char *arg = (char *)argv[0];
memset(&g_ctx, 0, sizeof(g_ctx));
if ((arg && (!strcmp(arg, "-h") || !strcmp(arg, "--help")))) {
printHelp(argv[0]);
return -1;
}
if (argc == 1) {
// Run using default arguments
g_ctx.input_nv12_file = sdkFindFilePath("test1920x1080.nv12", argv[0]);
if (g_ctx.input_nv12_file == NULL) {
printf("Cannot find input file test1920x1080.nv12\n Exiting\n");
return EXIT_FAILURE;
}
g_ctx.width = 1920;
g_ctx.height = 1080;
g_ctx.dst_width = 640;
g_ctx.dst_height = 480;
g_ctx.batch = 24;
} else if (argc > 1) {
if (checkCmdLineFlag(argc, (const char **)argv, "width")) {
g_ctx.width = getCmdLineArgumentInt(argc, (const char **)argv, "width");
}
if (checkCmdLineFlag(argc, (const char **)argv, "height")) {
g_ctx.height = getCmdLineArgumentInt(argc, (const char **)argv, "height");
}
if (checkCmdLineFlag(argc, (const char **)argv, "pitch")) {
g_ctx.pitch = getCmdLineArgumentInt(argc, (const char **)argv, "pitch");
}
if (checkCmdLineFlag(argc, (const char **)argv, "input")) {
getCmdLineArgumentString(argc, (const char **)argv, "input",
(char **)&g_ctx.input_nv12_file);
}
if (checkCmdLineFlag(argc, (const char **)argv, "dst_width")) {
g_ctx.dst_width =
getCmdLineArgumentInt(argc, (const char **)argv, "dst_width");
}
if (checkCmdLineFlag(argc, (const char **)argv, "dst_height")) {
g_ctx.dst_height =
getCmdLineArgumentInt(argc, (const char **)argv, "dst_height");
}
if (checkCmdLineFlag(argc, (const char **)argv, "dst_pitch")) {
g_ctx.dst_pitch =
getCmdLineArgumentInt(argc, (const char **)argv, "dst_pitch");
}
if (checkCmdLineFlag(argc, (const char **)argv, "batch")) {
g_ctx.batch = getCmdLineArgumentInt(argc, (const char **)argv, "batch");
}
}
g_ctx.device = findCudaDevice(argc, (const char **)argv);
if ((g_ctx.width == 0) || (g_ctx.height == 0) || (g_ctx.dst_width == 0) ||
(g_ctx.dst_height == 0) || !g_ctx.input_nv12_file) {
printHelp(argv[0]);
return -1;
}
if (g_ctx.pitch == 0) g_ctx.pitch = g_ctx.width;
if (g_ctx.dst_pitch == 0) g_ctx.dst_pitch = g_ctx.dst_width;
return 0;
}
/*
load nv12 yuvfile data into GPU device memory with batch of copy
*/
static int loadNV12Frame(unsigned char *d_inputNV12) {
unsigned char *pNV12FrameData;
unsigned char *d_nv12;
int frameSize;
std::ifstream nv12File(g_ctx.input_nv12_file, std::ifstream::in | std::ios::binary);
if (!nv12File.is_open()) {
std::cerr << "Can't open files\n";
return -1;
}
frameSize = g_ctx.pitch * g_ctx.ctx_heights;
#if USE_UVM_MEM
pNV12FrameData = d_inputNV12;
#else
pNV12FrameData = (unsigned char *)malloc(frameSize);
if (pNV12FrameData == NULL) {
std::cerr << "Failed to malloc pNV12FrameData\n";
return -1;
}
#endif
nv12File.read((char *)pNV12FrameData, frameSize);
if (nv12File.gcount() < frameSize) {
std::cerr << "can't get one frame!\n";
return -1;
}
#if USE_UVM_MEM
// Prefetch to GPU for following GPU operation
cudaStreamAttachMemAsync(NULL, pNV12FrameData, 0, cudaMemAttachGlobal);
#endif
// expand one frame to multi frames for batch processing
d_nv12 = d_inputNV12;
for (int i = 0; i < g_ctx.batch; i++) {
checkCudaErrors(cudaMemcpy2D((void *)d_nv12, g_ctx.ctx_pitch,
pNV12FrameData, g_ctx.width, g_ctx.width,
g_ctx.ctx_heights, cudaMemcpyHostToDevice));
d_nv12 += g_ctx.ctx_pitch * g_ctx.ctx_heights;
}
#if (USE_UVM_MEM == 0)
free(pNV12FrameData);
#endif
nv12File.close();
return 0;
}
/*
1. resize interlace nv12 to target size
2. convert nv12 to bgr 3 progressive planars
*/
void nv12ResizeAndNV12ToBGR(unsigned char *d_inputNV12) {
unsigned char *d_resizedNV12;
float *d_outputBGR;
int size;
char filename[40];
/* allocate device memory for resized nv12 output */
size = g_ctx.dst_width * ceil(g_ctx.dst_height * 3.0f / 2.0f) * g_ctx.batch *
sizeof(unsigned char);
checkCudaErrors(cudaMalloc((void **)&d_resizedNV12, size));
/* allocate device memory for bgr output */
size = g_ctx.dst_pitch * g_ctx.dst_height * 3 * g_ctx.batch * sizeof(float);
checkCudaErrors(cudaMalloc((void **)&d_outputBGR, size));
cudaStream_t stream;
checkCudaErrors(cudaStreamCreate(&stream));
/* create cuda event handles */
cudaEvent_t start, stop;
checkCudaErrors(cudaEventCreate(&start));
checkCudaErrors(cudaEventCreate(&stop));
float elapsedTime = 0.0f;
/* resize interlace nv12 */
cudaEventRecord(start, 0);
for (int i = 0; i < TEST_LOOP; i++) {
resizeNV12Batch(d_inputNV12, g_ctx.ctx_pitch, g_ctx.width, g_ctx.height,
d_resizedNV12, g_ctx.dst_width, g_ctx.dst_width,
g_ctx.dst_height, g_ctx.batch);
}
cudaEventRecord(stop, 0);
cudaEventSynchronize(stop);
cudaEventElapsedTime(&elapsedTime, start, stop);
printf(
" CUDA resize nv12(%dx%d --> %dx%d), batch: %d,"
" average time: %.3f ms ==> %.3f ms/frame\n",
g_ctx.width, g_ctx.height, g_ctx.dst_width, g_ctx.dst_height, g_ctx.batch,
(elapsedTime / (TEST_LOOP * 1.0f)),
(elapsedTime / (TEST_LOOP * 1.0f)) / g_ctx.batch);
sprintf(filename, "resized_nv12_%dx%d", g_ctx.dst_width, g_ctx.dst_height);
/* convert nv12 to bgr 3 progressive planars */
cudaEventRecord(start, 0);
for (int i = 0; i < TEST_LOOP; i++) {
nv12ToBGRplanarBatch(d_resizedNV12, g_ctx.dst_pitch, // intput
d_outputBGR,
g_ctx.dst_pitch * sizeof(float), // output
g_ctx.dst_width, g_ctx.dst_height, // output
g_ctx.batch, 0);
}
cudaEventRecord(stop, 0);
cudaEventSynchronize(stop);
cudaEventElapsedTime(&elapsedTime, start, stop);
printf(
" CUDA convert nv12(%dx%d) to bgr(%dx%d), batch: %d,"
" average time: %.3f ms ==> %.3f ms/frame\n",
g_ctx.dst_width, g_ctx.dst_height, g_ctx.dst_width, g_ctx.dst_height,
g_ctx.batch, (elapsedTime / (TEST_LOOP * 1.0f)),
(elapsedTime / (TEST_LOOP * 1.0f)) / g_ctx.batch);
sprintf(filename, "converted_bgr_%dx%d", g_ctx.dst_width, g_ctx.dst_height);
dumpBGR(d_outputBGR, g_ctx.dst_pitch, g_ctx.dst_width, g_ctx.dst_height,
g_ctx.batch, (char *)"t1", filename);
/* release resources */
checkCudaErrors(cudaEventDestroy(start));
checkCudaErrors(cudaEventDestroy(stop));
checkCudaErrors(cudaStreamDestroy(stream));
checkCudaErrors(cudaFree(d_resizedNV12));
checkCudaErrors(cudaFree(d_outputBGR));
}
/*
1. convert nv12 to bgr 3 progressive planars
2. resize bgr 3 planars to target size
*/
void nv12ToBGRandBGRresize(unsigned char *d_inputNV12) {
float *d_bgr;
float *d_resizedBGR;
int size;
char filename[40];
/* allocate device memory for bgr output */
size = g_ctx.ctx_pitch * g_ctx.height * 3 * g_ctx.batch * sizeof(float);
checkCudaErrors(cudaMalloc((void **)&d_bgr, size));
/* allocate device memory for resized bgr output */
size = g_ctx.dst_width * g_ctx.dst_height * 3 * g_ctx.batch * sizeof(float);
checkCudaErrors(cudaMalloc((void **)&d_resizedBGR, size));
cudaStream_t stream;
checkCudaErrors(cudaStreamCreate(&stream));
/* create cuda event handles */
cudaEvent_t start, stop;
checkCudaErrors(cudaEventCreate(&start));
checkCudaErrors(cudaEventCreate(&stop));
float elapsedTime = 0.0f;
/* convert interlace nv12 to bgr 3 progressive planars */
cudaEventRecord(start, 0);
cudaDeviceSynchronize();
for (int i = 0; i < TEST_LOOP; i++) {
nv12ToBGRplanarBatch(d_inputNV12, g_ctx.ctx_pitch, d_bgr,
g_ctx.ctx_pitch * sizeof(float), g_ctx.width,
g_ctx.height, g_ctx.batch, 0);
}
cudaEventRecord(stop, 0);
cudaEventSynchronize(stop);
cudaEventElapsedTime(&elapsedTime, start, stop);
printf(
" CUDA convert nv12(%dx%d) to bgr(%dx%d), batch: %d,"
" average time: %.3f ms ==> %.3f ms/frame\n",
g_ctx.width, g_ctx.height, g_ctx.width, g_ctx.height, g_ctx.batch,
(elapsedTime / (TEST_LOOP * 1.0f)),
(elapsedTime / (TEST_LOOP * 1.0f)) / g_ctx.batch);
sprintf(filename, "converted_bgr_%dx%d", g_ctx.width, g_ctx.height);
/* resize bgr 3 progressive planars */
cudaEventRecord(start, 0);
for (int i = 0; i < TEST_LOOP; i++) {
resizeBGRplanarBatch(d_bgr, g_ctx.ctx_pitch, g_ctx.width, g_ctx.height,
d_resizedBGR, g_ctx.dst_width, g_ctx.dst_width,
g_ctx.dst_height, g_ctx.batch);
}
cudaEventRecord(stop, 0);
cudaEventSynchronize(stop);
cudaEventElapsedTime(&elapsedTime, start, stop);
printf(
" CUDA resize bgr(%dx%d --> %dx%d), batch: %d,"
" average time: %.3f ms ==> %.3f ms/frame\n",
g_ctx.width, g_ctx.height, g_ctx.dst_width, g_ctx.dst_height, g_ctx.batch,
(elapsedTime / (TEST_LOOP * 1.0f)),
(elapsedTime / (TEST_LOOP * 1.0f)) / g_ctx.batch);
memset(filename, 0, sizeof(filename));
sprintf(filename, "resized_bgr_%dx%d", g_ctx.dst_width, g_ctx.dst_height);
dumpBGR(d_resizedBGR, g_ctx.dst_pitch, g_ctx.dst_width, g_ctx.dst_height,
g_ctx.batch, (char *)"t2", filename);
/* release resources */
checkCudaErrors(cudaEventDestroy(start));
checkCudaErrors(cudaEventDestroy(stop));
checkCudaErrors(cudaStreamDestroy(stream));
checkCudaErrors(cudaFree(d_bgr));
checkCudaErrors(cudaFree(d_resizedBGR));
}
int main(int argc, char *argv[]) {
unsigned char *d_inputNV12;
if (parseCmdLine(argc, argv) < 0) return EXIT_FAILURE;
g_ctx.ctx_pitch = g_ctx.width;
int ctx_alignment = 32;
g_ctx.ctx_pitch += (g_ctx.ctx_pitch % ctx_alignment != 0)
? (ctx_alignment - g_ctx.ctx_pitch % ctx_alignment)
: 0;
g_ctx.ctx_heights = ceil(g_ctx.height * 3.0f / 2.0f);
/* load nv12 yuv data into d_inputNV12 with batch of copies */
#if USE_UVM_MEM
checkCudaErrors(cudaMallocManaged(
(void **)&d_inputNV12,
(g_ctx.ctx_pitch * g_ctx.ctx_heights * g_ctx.batch), cudaMemAttachHost));
printf("\nUSE_UVM_MEM\n");
#else
checkCudaErrors(
cudaMalloc((void **)&d_inputNV12,
(g_ctx.ctx_pitch * g_ctx.ctx_heights * g_ctx.batch)));
#endif
if (loadNV12Frame(d_inputNV12)) {
std::cerr << "failed to load batch data!\n";
return EXIT_FAILURE;
}
/* firstly resize nv12, then convert nv12 to bgr */
printf("\nTEST#1:\n");
nv12ResizeAndNV12ToBGR(d_inputNV12);
/* first convert nv12 to bgr, then resize bgr */
printf("\nTEST#2:\n");
nv12ToBGRandBGRresize(d_inputNV12);
checkCudaErrors(cudaFree(d_inputNV12));
return EXIT_SUCCESS;
}

View File

@@ -0,0 +1,152 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <stdlib.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <fstream>
#include <iostream>
#include <cuda.h>
#include <cuda_runtime.h>
#include "resize_convert.h"
#include "utils.h"
__global__ void floatToChar(float *src, unsigned char *dst, int height,
int width, int batchSize) {
int x = threadIdx.x + blockIdx.x * blockDim.x;
if (x >= height * width) return;
int offset = height * width * 3;
for (int j = 0; j < batchSize; j++) {
// b
*(dst + j * offset + x * 3 + 0) =
(unsigned char)*(src + j * offset + height * width * 0 + x);
// g
*(dst + j * offset + x * 3 + 1) =
(unsigned char)*(src + j * offset + height * width * 1 + x);
// r
*(dst + j * offset + x * 3 + 2) =
(unsigned char)*(src + j * offset + height * width * 2 + x);
}
}
void floatPlanarToChar(float *src, unsigned char *dst, int height, int width,
int batchSize) {
floatToChar<<<(height * width - 1) / 1024 + 1, 1024, 0, NULL>>>(
src, dst, height, width, batchSize);
}
void dumpRawBGR(float *d_srcBGR, int pitch, int width, int height,
int batchSize, char *folder, char *tag) {
float *bgr, *d_bgr;
int frameSize;
char directory[120];
char mkdir_cmd[256];
#if !defined(_WIN32)
sprintf(directory, "output/%s", folder);
sprintf(mkdir_cmd, "mkdir -p %s 2> /dev/null", directory);
#else
sprintf(directory, "output\\%s", folder);
sprintf(mkdir_cmd, "mkdir %s 2> nul", directory);
#endif
int ret = system(mkdir_cmd);
frameSize = width * height * 3 * sizeof(float);
bgr = (float *)malloc(frameSize);
if (bgr == NULL) {
std::cerr << "Failed malloc for bgr\n";
return;
}
d_bgr = d_srcBGR;
for (int i = 0; i < batchSize; i++) {
char filename[120];
std::ofstream *outputFile;
checkCudaErrors(cudaMemcpy((void *)bgr, (void *)d_bgr, frameSize,
cudaMemcpyDeviceToHost));
sprintf(filename, "%s/%s_%d.raw", directory, tag, (i + 1));
outputFile = new std::ofstream(filename);
if (outputFile) {
outputFile->write((char *)bgr, frameSize);
delete outputFile;
}
d_bgr += pitch * height * 3;
}
free(bgr);
}
void dumpBGR(float *d_srcBGR, int pitch, int width, int height, int batchSize,
char *folder, char *tag) {
dumpRawBGR(d_srcBGR, pitch, width, height, batchSize, folder, tag);
}
void dumpYUV(unsigned char *d_nv12, int size, char *folder, char *tag) {
unsigned char *nv12Data;
std::ofstream *nv12File;
char filename[120];
char directory[60];
char mkdir_cmd[256];
#if !defined(_WIN32)
sprintf(directory, "output/%s", folder);
sprintf(mkdir_cmd, "mkdir -p %s 2> /dev/null", directory);
#else
sprintf(directory, "output\\%s", folder);
sprintf(mkdir_cmd, "mkdir %s 2> nul", directory);
#endif
int ret = system(mkdir_cmd);
sprintf(filename, "%s/%s.nv12", directory, tag);
nv12File = new std::ofstream(filename);
if (nv12File == NULL) {
std::cerr << "Failed to new " << filename;
return;
}
nv12Data = (unsigned char *)malloc(size * (sizeof(char)));
if (nv12Data == NULL) {
std::cerr << "Failed to allcoate memory\n";
return;
}
cudaMemcpy((void *)nv12Data, (void *)d_nv12, size, cudaMemcpyDeviceToHost);
nv12File->write((const char *)nv12Data, size);
free(nv12Data);
delete nv12File;
}

View File

@@ -0,0 +1,37 @@
/* Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __H_UTIL_
#define __H_UTIL_
extern "C"
void dumpBGR(float *d_srcBGR, int pitch, int width, int height,
int batchSize, char *folder, char *tag);
extern "C"
void dumpYUV(unsigned char *d_nv12, int size, char *folder, char *tag);
#endif

View File

@@ -0,0 +1,349 @@
################################################################################
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# * Neither the name of NVIDIA CORPORATION nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
# OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
################################################################################
#
# Makefile project only supported on Mac OS X and Linux Platforms)
#
################################################################################
# Location of the CUDA Toolkit
CUDA_PATH ?= /usr/local/cuda
##############################
# start deprecated interface #
##############################
ifeq ($(x86_64),1)
$(info WARNING - x86_64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=x86_64 instead)
TARGET_ARCH ?= x86_64
endif
ifeq ($(ARMv7),1)
$(info WARNING - ARMv7 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=armv7l instead)
TARGET_ARCH ?= armv7l
endif
ifeq ($(aarch64),1)
$(info WARNING - aarch64 variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=aarch64 instead)
TARGET_ARCH ?= aarch64
endif
ifeq ($(ppc64le),1)
$(info WARNING - ppc64le variable has been deprecated)
$(info WARNING - please use TARGET_ARCH=ppc64le instead)
TARGET_ARCH ?= ppc64le
endif
ifneq ($(GCC),)
$(info WARNING - GCC variable has been deprecated)
$(info WARNING - please use HOST_COMPILER=$(GCC) instead)
HOST_COMPILER ?= $(GCC)
endif
ifneq ($(abi),)
$(error ERROR - abi variable has been removed)
endif
############################
# end deprecated interface #
############################
# architecture
HOST_ARCH := $(shell uname -m)
TARGET_ARCH ?= $(HOST_ARCH)
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le armv7l))
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifneq (,$(filter $(TARGET_ARCH),x86_64 aarch64 sbsa ppc64le))
TARGET_SIZE := 64
else ifneq (,$(filter $(TARGET_ARCH),armv7l))
TARGET_SIZE := 32
endif
else
TARGET_SIZE := $(shell getconf LONG_BIT)
endif
else
$(error ERROR - unsupported value $(TARGET_ARCH) for TARGET_ARCH!)
endif
# sbsa and aarch64 systems look similar. Need to differentiate them at host level for now.
ifeq ($(HOST_ARCH),aarch64)
ifeq ($(CUDA_PATH)/targets/sbsa-linux,$(shell ls -1d $(CUDA_PATH)/targets/sbsa-linux 2>/dev/null))
HOST_ARCH := sbsa
TARGET_ARCH := sbsa
endif
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq (,$(filter $(HOST_ARCH)-$(TARGET_ARCH),aarch64-armv7l x86_64-armv7l x86_64-aarch64 x86_64-sbsa x86_64-ppc64le))
$(error ERROR - cross compiling from $(HOST_ARCH) to $(TARGET_ARCH) is not supported!)
endif
endif
# When on native aarch64 system with userspace of 32-bit, change TARGET_ARCH to armv7l
ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_SIZE),aarch64-aarch64-32)
TARGET_ARCH = armv7l
endif
# operating system
HOST_OS := $(shell uname -s 2>/dev/null | tr "[:upper:]" "[:lower:]")
TARGET_OS ?= $(HOST_OS)
ifeq (,$(filter $(TARGET_OS),linux darwin qnx android))
$(error ERROR - unsupported value $(TARGET_OS) for TARGET_OS!)
endif
# host compiler
ifeq ($(TARGET_OS),darwin)
ifeq ($(shell expr `xcodebuild -version | grep -i xcode | awk '{print $$2}' | cut -d'.' -f1` \>= 5),1)
HOST_COMPILER ?= clang++
endif
else ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(HOST_ARCH)-$(TARGET_ARCH),x86_64-armv7l)
ifeq ($(TARGET_OS),linux)
HOST_COMPILER ?= arm-linux-gnueabihf-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/arm-unknown-nto-qnx6.6.0eabi-g++
else ifeq ($(TARGET_OS),android)
HOST_COMPILER ?= arm-linux-androideabi-g++
endif
else ifeq ($(TARGET_ARCH),aarch64)
ifeq ($(TARGET_OS), linux)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_OS),qnx)
ifeq ($(QNX_HOST),)
$(error ERROR - QNX_HOST must be passed to the QNX host toolchain)
endif
ifeq ($(QNX_TARGET),)
$(error ERROR - QNX_TARGET must be passed to the QNX target toolchain)
endif
export QNX_HOST
export QNX_TARGET
HOST_COMPILER ?= $(QNX_HOST)/usr/bin/q++
else ifeq ($(TARGET_OS), android)
HOST_COMPILER ?= aarch64-linux-android-clang++
endif
else ifeq ($(TARGET_ARCH),sbsa)
HOST_COMPILER ?= aarch64-linux-gnu-g++
else ifeq ($(TARGET_ARCH),ppc64le)
HOST_COMPILER ?= powerpc64le-linux-gnu-g++
endif
endif
HOST_COMPILER ?= g++
NVCC := $(CUDA_PATH)/bin/nvcc -ccbin $(HOST_COMPILER)
# internal flags
NVCCFLAGS := -m${TARGET_SIZE}
CCFLAGS :=
LDFLAGS :=
# build flags
ifeq ($(TARGET_OS),darwin)
LDFLAGS += -rpath $(CUDA_PATH)/lib
CCFLAGS += -arch $(HOST_ARCH)
else ifeq ($(HOST_ARCH)-$(TARGET_ARCH)-$(TARGET_OS),x86_64-armv7l-linux)
LDFLAGS += --dynamic-linker=/lib/ld-linux-armhf.so.3
CCFLAGS += -mfloat-abi=hard
else ifeq ($(TARGET_OS),android)
LDFLAGS += -pie
CCFLAGS += -fpie -fpic -fexceptions
endif
ifneq ($(TARGET_ARCH),$(HOST_ARCH))
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/arm-linux-gnueabihf
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
ifneq ($(TARGET_FS),)
GCCVERSIONLTEQ46 := $(shell expr `$(HOST_COMPILER) -dumpversion` \<= 4.6)
ifeq ($(GCCVERSIONLTEQ46),1)
CCFLAGS += --sysroot=$(TARGET_FS)
endif
LDFLAGS += --sysroot=$(TARGET_FS)
LDFLAGS += -rpath-link=$(TARGET_FS)/lib -L$(TARGET_FS)/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/lib/aarch64-linux-gnu -L$(TARGET_FS)/lib/aarch64-linux-gnu
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib -L$(TARGET_FS)/usr/lib
LDFLAGS += -rpath-link=$(TARGET_FS)/usr/lib/aarch64-linux-gnu -L$(TARGET_FS)/usr/lib/aarch64-linux-gnu
LDFLAGS += --unresolved-symbols=ignore-in-shared-libs
CCFLAGS += -isystem=$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include -I$(TARGET_FS)/usr/include/libdrm
CCFLAGS += -isystem=$(TARGET_FS)/usr/include/aarch64-linux-gnu -I$(TARGET_FS)/usr/include/aarch64-linux-gnu
endif
endif
ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
NVCCFLAGS += --qpp-config 5.4.0,gcc_ntoaarch64le
CCFLAGS += -DWIN_INTERFACE_CUSTOM -I/usr/include/aarch64-qnx-gnu
LDFLAGS += -lsocket
LDFLAGS += -L/usr/lib/aarch64-qnx-gnu
CCFLAGS += "-Wl\,-rpath-link\,/usr/lib/aarch64-qnx-gnu"
ifdef TARGET_OVERRIDE
LDFLAGS += -lslog2
endif
ifneq ($(TARGET_FS),)
LDFLAGS += -L$(TARGET_FS)/usr/lib
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/lib"
LDFLAGS += -L$(TARGET_FS)/usr/libnvidia
CCFLAGS += "-Wl\,-rpath-link\,$(TARGET_FS)/usr/libnvidia"
CCFLAGS += -I$(TARGET_FS)/../include
endif
endif
endif
ifdef TARGET_OVERRIDE # cuda toolkit targets override
NVCCFLAGS += -target-dir $(TARGET_OVERRIDE)
endif
# Install directory of different arch
CUDA_INSTALL_TARGET_DIR :=
ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-linux)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-gnueabihf/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-linux)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),sbsa-linux)
CUDA_INSTALL_TARGET_DIR = targets/sbsa-linux/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-android)
CUDA_INSTALL_TARGET_DIR = targets/armv7-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-android)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-linux-androideabi/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),armv7l-qnx)
CUDA_INSTALL_TARGET_DIR = targets/ARMv7-linux-QNX/
else ifeq ($(TARGET_ARCH)-$(TARGET_OS),aarch64-qnx)
CUDA_INSTALL_TARGET_DIR = targets/aarch64-qnx/
else ifeq ($(TARGET_ARCH),ppc64le)
CUDA_INSTALL_TARGET_DIR = targets/ppc64le-linux/
endif
# Debug build flags
ifeq ($(dbg),1)
NVCCFLAGS += -g -G
BUILD_TYPE := debug
else
BUILD_TYPE := release
endif
ALL_CCFLAGS :=
ALL_CCFLAGS += $(NVCCFLAGS)
ALL_CCFLAGS += $(EXTRA_NVCCFLAGS)
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(CCFLAGS))
ALL_CCFLAGS += $(addprefix -Xcompiler ,$(EXTRA_CCFLAGS))
SAMPLE_ENABLED := 1
# This sample is not supported on QNX
ifeq ($(TARGET_OS),qnx)
$(info >>> WARNING - UnifiedMemoryPerf is not supported on QNX - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ALL_LDFLAGS :=
ALL_LDFLAGS += $(ALL_CCFLAGS)
ALL_LDFLAGS += $(addprefix -Xlinker ,$(LDFLAGS))
ALL_LDFLAGS += $(addprefix -Xlinker ,$(EXTRA_LDFLAGS))
# Common includes and paths for CUDA
INCLUDES := -I../../Common
LIBRARIES :=
################################################################################
# Gencode arguments
ifeq ($(TARGET_ARCH),$(filter $(TARGET_ARCH),armv7l aarch64))
SMS ?= 35 37 50 52 60 61 70 72 75 80 86
else
SMS ?= 35 37 50 52 60 61 70 75 80 86
endif
ifeq ($(SMS),)
$(info >>> WARNING - no SM architectures have been specified - waiving sample <<<)
SAMPLE_ENABLED := 0
endif
ifeq ($(GENCODE_FLAGS),)
# Generate SASS code for each SM architecture listed in $(SMS)
$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
# Generate PTX code from the highest SM architecture in $(SMS) to guarantee forward-compatibility
HIGHEST_SM := $(lastword $(sort $(SMS)))
ifneq ($(HIGHEST_SM),)
GENCODE_FLAGS += -gencode arch=compute_$(HIGHEST_SM),code=compute_$(HIGHEST_SM)
endif
endif
ALL_CCFLAGS += --threads 0
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: UnifiedMemoryPerf
check.deps:
ifeq ($(SAMPLE_ENABLED),0)
@echo "Sample will be waived due to the above missing dependencies"
else
@echo "Sample is ready - all dependencies have been met"
endif
commonKernels.o:commonKernels.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
helperFunctions.o:helperFunctions.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
matrixMultiplyPerf.o:matrixMultiplyPerf.cu
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
UnifiedMemoryPerf: commonKernels.o helperFunctions.o matrixMultiplyPerf.o
$(EXEC) $(NVCC) $(ALL_LDFLAGS) $(GENCODE_FLAGS) -o $@ $+ $(LIBRARIES)
$(EXEC) mkdir -p ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
$(EXEC) cp $@ ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)
run: build
$(EXEC) ./UnifiedMemoryPerf
clean:
rm -f UnifiedMemoryPerf commonKernels.o helperFunctions.o matrixMultiplyPerf.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/UnifiedMemoryPerf
clobber: clean

View File

@@ -0,0 +1,75 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>UnifiedMemoryPerf</name>
<cuda_api_list>
<toolkit>cudaMallocManaged</toolkit>
<toolkit>cudaStreamAttachMemAsync</toolkit>
<toolkit>cudaMemcpyAsync</toolkit>
<toolkit>cudaMallocHost</toolkit>
<toolkit>cudaMalloc</toolkit>
</cuda_api_list>
<description><![CDATA[This sample demonstrates the performance comparision using matrix multiplication kernel of Unified Memory with/without hints and other types of memory like zero copy buffers, pageable, pagelocked memory performing synchronous and Asynchronous transfers on a single GPU.]]></description>
<devicecompilation>whole</devicecompilation>
<includepaths>
<path>./</path>
<path>../</path>
<path>../../common/inc</path>
</includepaths>
<keyconcepts>
<concept level="basic">CUDA Systems Integration</concept>
<concept level="basic">Unified Memory</concept>
<concept level="basic">CUDA Streams and Events</concept>
<concept level="basic">Pinned System Paged Memory</concept>
</keyconcepts>
<keywords>
<keyword>CUDA</keyword>
<keyword>Unified Memory</keyword>
<keyword>Pinned Memory</keyword>
<keyword>Zero copy buffer</keyword>
<keyword>UVM</keyword>
<keyword>Streams</keyword>
</keywords>
<libraries>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>matrixMultiplyPerf.cu</primary_file>
<required_dependencies>
<dependency>UVM</dependency>
</required_dependencies>
<scopes>
<scope>1:CUDA Basic Topics</scope>
<scope>1:CUDA Systems Integration</scope>
<scope>1:Unified Memory</scope>
</scopes>
<supported_envs>
<env>
<arch>x86_64</arch>
<platform>linux</platform>
</env>
<env>
<arch>x86_64</arch>
<platform>macosx</platform>
</env>
<env>
<platform>windows7</platform>
</env>
<env>
<arch>arm</arch>
</env>
<env>
<arch>aarch64</arch>
</env>
<env>
<arch>ppc64le</arch>
<platform>linux</platform>
</env>
</supported_envs>
<supported_sm_architectures>
<from>3.0</from>
</supported_sm_architectures>
<title>Unified and other CUDA Memories Performance</title>
<type>exe</type>
</entry>

View File

@@ -0,0 +1,72 @@
# UnifiedMemoryPerf - Unified and other CUDA Memories Performance
## Description
This sample demonstrates the performance comparision using matrix multiplication kernel of Unified Memory with/without hints and other types of memory like zero copy buffers, pageable, pagelocked memory performing synchronous and Asynchronous transfers on a single GPU.
## Key Concepts
CUDA Systems Integration, Unified Memory, CUDA Streams and Events, Pinned System Paged Memory
## Supported SM Architectures
## Supported OSes
Linux, Windows
## Supported CPU Architecture
x86_64, ppc64le, armv7l, aarch64
## CUDA APIs involved
### [CUDA Runtime API](http://docs.nvidia.com/cuda/cuda-runtime-api/index.html)
cudaMallocManaged, cudaStreamAttachMemAsync, cudaMemcpyAsync, cudaMallocHost, cudaMalloc
## Dependencies needed to build/run
[UVM](../../README.md#uvm)
## Prerequisites
Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run
### Windows
The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format:
```
*_vs<version>.sln - for Visual Studio <version>
```
Each individual sample has its own set of solution files in its directory:
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.
> **Note:** Some samples require that the Microsoft DirectX SDK (June 2010 or newer) be installed and that the VC++ directory paths are properly set up (**Tools > Options...**). Check DirectX Dependencies section for details."
### Linux
The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make:
```
$ cd <sample_dir>
$ make
```
The samples makefiles can take advantage of certain options:
* **TARGET_ARCH=<arch>** - cross-compile targeting a specific architecture. Allowed architectures are x86_64, ppc64le, armv7l, aarch64.
By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equivalent of setting TARGET_ARCH=x86_64.<br/>
`$ make TARGET_ARCH=x86_64` <br/> `$ make TARGET_ARCH=ppc64le` <br/> `$ make TARGET_ARCH=armv7l` <br/> `$ make TARGET_ARCH=aarch64` <br/>
See [here](http://docs.nvidia.com/cuda/cuda-samples/index.html#cross-samples) for more details.
* **dbg=1** - build with debug symbols
```
$ make dbg=1
```
* **SMS="A B ..."** - override the SM architectures for which the sample will be built, where `"A B ..."` is a space-delimited list of SM architectures. For example, to generate SASS for SM 50 and SM 60, use `SMS="50 60"`.
```
$ make SMS="50 60"
```
* **HOST_COMPILER=<host_compiler>** - override the default g++ host compiler. See the [Linux Installation Guide](http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements) for a list of supported host compilers.
```
$ make HOST_COMPILER=g++
```
## References (for more details)

View File

@@ -0,0 +1,20 @@

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "UnifiedMemoryPerf", "UnifiedMemoryPerf_vs2017.vcxproj", "{997E0757-EA74-4A4E-A0FC-47D8C8831A15}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|x64 = Debug|x64
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.ActiveCfg = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Debug|x64.Build.0 = Debug|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.ActiveCfg = Release|x64
{997E0757-EA74-4A4E-A0FC-47D8C8831A15}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal

View File

@@ -0,0 +1,116 @@
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup>
<CUDAPropsPath Condition="'$(CUDAPropsPath)'==''">$(VCTargetsPath)\BuildCustomizations</CUDAPropsPath>
</PropertyGroup>
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{997E0757-EA74-4A4E-A0FC-47D8C8831A15}</ProjectGuid>
<RootNamespace>UnifiedMemoryPerf_vs2017</RootNamespace>
<ProjectName>UnifiedMemoryPerf</ProjectName>
<CudaToolkitCustomDir />
</PropertyGroup>
<PropertyGroup Condition="'$(WindowsTargetPlatformVersion)'==''">
<LatestTargetPlatformVersion>$([Microsoft.Build.Utilities.ToolLocationHelper]::GetLatestSDKTargetPlatformVersion('Windows', '10.0'))</LatestTargetPlatformVersion>
<WindowsTargetPlatformVersion Condition="'$(WindowsTargetPlatformVersion)' == ''">$(LatestTargetPlatformVersion)</WindowsTargetPlatformVersion>
<TargetPlatformVersion>$(WindowsTargetPlatformVersion)</TargetPlatformVersion>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup>
<ConfigurationType>Application</ConfigurationType>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v141</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'">
<UseDebugLibraries>true</UseDebugLibraries>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Release'">
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup>
<IntDir>$(Platform)/$(Configuration)/</IntDir>
<IncludePath>$(IncludePath)</IncludePath>
<CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
<CodeAnalysisRules />
<CodeAnalysisRuleAssemblies />
</PropertyGroup>
<PropertyGroup Condition="'$(Platform)'=='x64'">
<OutDir>../../bin/win64/$(Configuration)/</OutDir>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<WarningLevel>Level3</WarningLevel>
<PreprocessorDefinitions>WIN32;_MBCS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>./;$(CudaToolkitDir)/include;../../Common;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/UnifiedMemoryPerf.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,sm_35;compute_37,sm_37;compute_50,sm_50;compute_52,sm_52;compute_60,sm_60;compute_61,sm_61;compute_70,sm_70;compute_75,sm_75;compute_80,sm_80;compute_86,sm_86;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Debug'">
<ClCompile>
<Optimization>Disabled</Optimization>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<LinkTimeCodeGeneration>Default</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MTd</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemGroup>
<CudaCompile Include="commonKernels.cu" />
<ClCompile Include="helperFunctions.cpp" />
<CudaCompile Include="matrixMultiplyPerf.cu" />
<ClInclude Include="commonDefs.hpp" />
<ClInclude Include="commonKernels.hpp" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

Some files were not shown because too many files have changed in this diff Show More