NASP/cuda-samples
85
0
Fork 0
mirror of https://github.com/NVIDIA/cuda-samples.git synced 2026-10-23 13:18:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

add and update samples for CUDA 11.5

Browse Source
This commit is contained in:
Rutwik Choughule
2021-10-21 16:34:49 +05:30
parent 3342d604fe
commit 1f76a2d110
2796 changed files with 1511725 additions and 827 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
Samples/FilterBorderControlNPP/.vscode/c_cpp_properties.json vendored Normal file
View File

@@ -0,0 +1,18 @@
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/../../Common"
],
"defines": [],
"compilerPath": "/usr/local/cuda/bin/nvcc",
"cStandard": "gnu17",
"cppStandard": "gnu++14",
"intelliSenseMode": "linux-gcc-x64",
"configurationProvider": "ms-vscode.makefile-tools"
}
],
"version": 4
}

7
Samples/FilterBorderControlNPP/.vscode/extensions.json vendored Normal file
View File

@@ -0,0 +1,7 @@
{
"recommendations": [
"nvidia.nsight-vscode-edition",
"ms-vscode.cpptools",
"ms-vscode.makefile-tools"
]
}

10
Samples/FilterBorderControlNPP/.vscode/launch.json vendored Normal file
View File

@@ -0,0 +1,10 @@
{
"configurations": [
{
"name": "CUDA C++: Launch",
"type": "cuda-gdb",
"request": "launch",
"program": "${workspaceFolder}/FilterBorderControlNPP"
}
]
}

15
Samples/FilterBorderControlNPP/.vscode/tasks.json vendored Normal file
View File

@@ -0,0 +1,15 @@
{
"version": "2.0.0",
"tasks": [
{
"label": "sample",
"type": "shell",
"command": "make dbg=1",
"problemMatcher": ["$nvcc"],
"group": {
"kind": "build",
"isDefault": true
}
}
]
}

502
Samples/FilterBorderControlNPP/FilterBorderControlNPP.cpp Normal file
View File

@@ -0,0 +1,502 @@
/* Copyright (c) 2021, 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.
*/
// USING NPP IMAGE FILTER SOURCE IMAGE BORDER CONTROL
// This sample demonstrates how any border version of an NPP filtering function
// can be used in the most common mode, with border control enabled. Mentioned
// functions can be used to duplicate the results of the equivalent non-border
// version of the NPP functions. They can be also used for enabling and
// disabling border control on various source image edges depending on what
// portion of the source image is being used as input.
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
#define WINDOWS_LEAN_AND_MEAN
#define NOMINMAX
#include <windows.h>
#pragma warning(disable : 4819)
#endif
#include <Exceptions.h>
#include <ImageIO.h>
#include <ImagesCPU.h>
#include <ImagesNPP.h>
#include <cuda_runtime.h>
#include <helper_cuda.h>
#include <helper_string.h>
#include <npp.h>
#include <string.h>
#include <fstream>
#include <iostream>
inline int cudaDeviceInit(int argc, const char **argv) {
int deviceCount;
checkCudaErrors(cudaGetDeviceCount(&deviceCount));
if (deviceCount == 0) {
std::cerr << "CUDA error: no devices supporting CUDA." << std::endl;
exit(EXIT_FAILURE);
}
int dev = findCudaDevice(argc, argv);
cudaDeviceProp deviceProp;
cudaGetDeviceProperties(&deviceProp, dev);
std::cerr << "cudaSetDevice GPU" << dev << " = " << deviceProp.name
<< std::endl;
checkCudaErrors(cudaSetDevice(dev));
return dev;
}
bool printfNPPinfo(int argc, char *argv[]) {
const NppLibraryVersion *libVer = nppGetLibVersion();
printf("NPP Library Version %d.%d.%d\n", libVer->major, libVer->minor,
libVer->build);
int driverVersion, runtimeVersion;
cudaDriverGetVersion(&driverVersion);
cudaRuntimeGetVersion(&runtimeVersion);
printf(" CUDA Driver Version: %d.%d\n", driverVersion / 1000,
(driverVersion % 100) / 10);
printf(" CUDA Runtime Version: %d.%d\n", runtimeVersion / 1000,
(runtimeVersion % 100) / 10);
// Min spec is SM 1.0 devices
bool bVal = checkCudaCapabilities(1, 0);
return bVal;
}
int main(int argc, char *argv[]) {
printf("%s Starting...\n\n", argv[0]);
try {
const char *inputFile = "Lena.pgm";
std::string sFilename = inputFile;
std::string sOutputDir = "./";
cudaDeviceInit(argc, (const char **)argv);
if (printfNPPinfo(argc, argv) == false) {
cudaDeviceReset();
exit(EXIT_SUCCESS);
}
char *filePath;
if (checkCmdLineFlag(argc, (const char **)argv, "input")) {
getCmdLineArgumentString(argc, (const char **)argv, "input", &filePath);
} else {
filePath = sdkFindFilePath(inputFile, argv[0]);
}
if (!filePath) {
std::cerr << "Couldn't find input file " << sFilename << std::endl;
exit(1);
}
sFilename = filePath;
// if we specify the filename at the command line, then we only test
// sFilename[0].
int file_errors = 0;
std::ifstream infile(sFilename.data(), std::ifstream::in);
if (infile.good()) {
std::cout << "gradientFilterBorderNPP opened <" << sFilename.data()
<< "> successfully!" << std::endl;
file_errors = 0;
infile.close();
} else {
std::cout << "gradientFilterBorderNPP unable to open <"
<< sFilename.data() << ">" << std::endl;
file_errors++;
infile.close();
}
if (file_errors > 0) {
cudaDeviceReset();
exit(EXIT_FAILURE);
}
std::string sResultBaseFilename = sFilename;
std::string::size_type dot = sResultBaseFilename.rfind('.');
if (dot != std::string::npos) {
sResultBaseFilename = sResultBaseFilename.substr(0, dot);
}
std::string sResultXFilename =
sOutputDir + sFilename + "_gradientVectorPrewittBorderX_Vertical.pgm";
std::string sResultYFilename = sResultBaseFilename;
// sResultXFilename += "_gradientVectorPrewittBorderX_Vertical.pgm";
sResultYFilename += "_gradientVectorPrewittBorderY_Horizontal.pgm";
// if (checkCmdLineFlag(argc, (const char **)argv, "output"))
// {
// char *outputFilePath;
// getCmdLineArgumentString(argc, (const char **)argv, "output",
// &outputFilePath); sResultBaseFilename = outputFilePath;
// }
// declare a host image object for an 8-bit grayscale image
npp::ImageCPU_8u_C1 oHostSrc;
// load gray-scale image from disk
npp::loadImage(sFilename, oHostSrc);
// declare a device image and copy construct from the host image,
// i.e. upload host to device
npp::ImageNPP_8u_C1 oDeviceSrc(oHostSrc);
NppiSize oSrcSize = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
NppiPoint oSrcOffset = {0, 0};
// create struct with ROI size
NppiSize oSizeROI = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
// allocate device destination images of appropriatedly size
npp::ImageNPP_16s_C1 oDeviceDstX(oSizeROI.width, oSizeROI.height);
npp::ImageNPP_16s_C1 oDeviceDstY(oSizeROI.width, oSizeROI.height);
// run Prewitt edge detection gradient vector filter
NPP_CHECK_NPP(nppiGradientVectorPrewittBorder_8u16s_C1R(
oDeviceSrc.data(), oDeviceSrc.pitch(), oSrcSize, oSrcOffset,
oDeviceDstX.data(), oDeviceDstX.pitch(), oDeviceDstY.data(),
oDeviceDstY.pitch(), 0, 0, 0, 0, oSizeROI, NPP_MASK_SIZE_3_X_3,
nppiNormL1, NPP_BORDER_REPLICATE));
// allocate device destination images of appropriatedly size
npp::ImageNPP_8u_C1 oDeviceDstOutX(oSizeROI.width, oSizeROI.height);
npp::ImageNPP_8u_C1 oDeviceDstOutY(oSizeROI.width, oSizeROI.height);
// convert 16s_C1 result images to binary 8u_C1 output images using constant
// value to adjust amount of visible detail
NPP_CHECK_NPP(nppiCompareC_16s_C1R(
oDeviceDstX.data(), oDeviceDstX.pitch(), 32, oDeviceDstOutX.data(),
oDeviceDstOutX.pitch(), oSizeROI, NPP_CMP_GREATER_EQ));
NPP_CHECK_NPP(nppiCompareC_16s_C1R(
oDeviceDstY.data(), oDeviceDstY.pitch(), 32, oDeviceDstOutY.data(),
oDeviceDstOutY.pitch(), oSizeROI, NPP_CMP_GREATER_EQ));
// create host images for the results
npp::ImageCPU_8u_C1 oHostDstX(oDeviceDstOutX.size());
npp::ImageCPU_8u_C1 oHostDstY(oDeviceDstOutY.size());
// and copy the device result data into them
oDeviceDstOutX.copyTo(oHostDstX.data(), oHostDstX.pitch());
oDeviceDstOutY.copyTo(oHostDstY.data(), oHostDstY.pitch());
saveImage(sResultXFilename, oHostDstX);
std::cout << "Saved image: " << sResultXFilename << std::endl;
saveImage(sResultYFilename, oHostDstY);
std::cout << "Saved image: " << sResultYFilename << std::endl;
// now use the Prewitt gradient border filter function in such a way that no
// border replication operations will be applied
// create a Prewitt filter mask size object, Prewitt uses a 3x3 filter
// kernel
NppiSize oMaskSize = {3, 3};
// create a size object for the enlarged source image
NppiSize oEnlargedSrcSize = {oSrcSize.width + oMaskSize.width - 1,
oSrcSize.height + oMaskSize.height - 1};
// create an enlarged device source image
npp::ImageNPP_8u_C1 oEnlargedDeviceSrc(oEnlargedSrcSize.width,
oEnlargedSrcSize.height);
// copy and enlarge the original device source image and surround it with a
// white edge (border)
NPP_CHECK_NPP(nppiCopyConstBorder_8u_C1R(
oDeviceSrc.data(), oDeviceSrc.pitch(), oSrcSize,
oEnlargedDeviceSrc.data(), oEnlargedDeviceSrc.pitch(), oEnlargedSrcSize,
oMaskSize.width / 2, oMaskSize.height / 2, 255));
// adjust oEnlargedDeviceSrc pixel pointer to point to the first pixel of
// the original source image in the enlarged source image
const Npp8u *pTemp =
reinterpret_cast<const Npp8u *>(oEnlargedDeviceSrc.data());
pTemp += (oMaskSize.height / 2) * oEnlargedDeviceSrc.pitch();
const Npp8u *pAdjustedSrc =
reinterpret_cast<const Npp8u *>((void *)(pTemp));
pAdjustedSrc += oMaskSize.width / 2;
// create device output images for the no source border results
npp::ImageNPP_8u_C1 oDeviceDstOutXNoBorders(oSizeROI.width,
oSizeROI.height);
npp::ImageNPP_8u_C1 oDeviceDstOutYNoBorders(oSizeROI.width,
oSizeROI.height);
// tell the filter function what cartesian pixel position pAdjustedSrc is
// pointing to within the enlarged source image
oSrcOffset.x += oMaskSize.width / 2;
oSrcOffset.y += oMaskSize.height / 2;
// run Prewitt edge detection gradient vector filter bypassing border
// control due to enlarged source image
NPP_CHECK_NPP(nppiGradientVectorPrewittBorder_8u16s_C1R(
pAdjustedSrc, oEnlargedDeviceSrc.pitch(), oEnlargedSrcSize, oSrcOffset,
oDeviceDstX.data(), oDeviceDstX.pitch(), oDeviceDstY.data(),
oDeviceDstY.pitch(), 0, 0, 0, 0, oSizeROI, NPP_MASK_SIZE_3_X_3,
nppiNormL1, NPP_BORDER_REPLICATE));
// convert 16s_C1 result images to binary 8u_C1 output images using constant
// value to adjust amount of visible detail
NPP_CHECK_NPP(nppiCompareC_16s_C1R(oDeviceDstX.data(), oDeviceDstX.pitch(),
32, oDeviceDstOutXNoBorders.data(),
oDeviceDstOutXNoBorders.pitch(),
oSizeROI, NPP_CMP_GREATER_EQ));
NPP_CHECK_NPP(nppiCompareC_16s_C1R(oDeviceDstY.data(), oDeviceDstY.pitch(),
32, oDeviceDstOutYNoBorders.data(),
oDeviceDstOutYNoBorders.pitch(),
oSizeROI, NPP_CMP_GREATER_EQ));
// create additional output files
std::string sResultXNoBordersFilename = sResultBaseFilename;
std::string sResultYNoBordersFilename = sResultBaseFilename;
sResultXNoBordersFilename +=
"_gradientVectorPrewittBorderX_Vertical_WithNoSourceBorders.pgm";
sResultYNoBordersFilename +=
"_gradientVectorPrewittBorderY_Horizontal_WithNoSourceBorders.pgm";
// copy the device result data into the host output images
oDeviceDstOutXNoBorders.copyTo(oHostDstX.data(), oHostDstX.pitch());
oDeviceDstOutYNoBorders.copyTo(oHostDstY.data(), oHostDstY.pitch());
saveImage(sResultXNoBordersFilename, oHostDstX);
std::cout << "Saved image: " << sResultXNoBordersFilename << std::endl;
saveImage(sResultYNoBordersFilename, oHostDstY);
std::cout << "Saved image: " << sResultYNoBordersFilename << std::endl;
// now diff the two output images, one using border control and one
// bypassing border control
// create device output images for the diff results
npp::ImageNPP_8u_C1 oDeviceDstOutXDiff(oSizeROI.width, oSizeROI.height);
npp::ImageNPP_8u_C1 oDeviceDstOutYDiff(oSizeROI.width, oSizeROI.height);
// diff the two 8u_C1 result images one with and one without border control
NPP_CHECK_NPP(nppiAbsDiff_8u_C1R(
oDeviceDstOutXNoBorders.data(), oDeviceDstOutXNoBorders.pitch(),
oDeviceDstOutX.data(), oDeviceDstOutX.pitch(),
oDeviceDstOutXDiff.data(), oDeviceDstOutXDiff.pitch(), oSizeROI));
NPP_CHECK_NPP(nppiAbsDiff_8u_C1R(
oDeviceDstOutYNoBorders.data(), oDeviceDstOutYNoBorders.pitch(),
oDeviceDstOutY.data(), oDeviceDstOutY.pitch(),
oDeviceDstOutYDiff.data(), oDeviceDstOutYDiff.pitch(), oSizeROI));
// create additional output files
std::string sResultXDiffFilename = sResultBaseFilename;
std::string sResultYDiffFilename = sResultBaseFilename;
sResultXDiffFilename +=
"_gradientVectorPrewittBorderX_Vertical_BorderDiffs.pgm";
sResultYDiffFilename +=
"_gradientVectorPrewittBorderY_Horizontal_BorderDiffs.pgm";
// copy the device result data into the host output images
oDeviceDstOutXDiff.copyTo(oHostDstX.data(), oHostDstX.pitch());
oDeviceDstOutYDiff.copyTo(oHostDstY.data(), oHostDstY.pitch());
saveImage(sResultXDiffFilename, oHostDstX);
std::cout << "Saved image: " << sResultXDiffFilename << std::endl;
saveImage(sResultYDiffFilename, oHostDstY);
std::cout << "Saved image: " << sResultYDiffFilename << std::endl;
// if you closely examine the above difference files (recommend using GIMP
// for viewing using scaling with no interpolation) you will see several
// single pixel differences (white pixels) along the right and bottom edges
// of the default vs. borderless images this happens because border pixels
// in the original source image are duplicated when the filter kernels
// overlap the edge of the source image when using the first version of the
// filter call but are actually sampled from the enlarged source image when
// using the second version of the filter call the technique used in the
// second filter call can be used with any filter border function in NPP to
// duplicate results that would be generated from a non-border filter
// function call by filling the border pixel outside the embedded source
// image with the appropriate border pixel values
// here is how to use border control to process a source image in multiple
// calls and get correct output in the destination image
// since the source image pointer already points to the beginning of the
// source image in the enlarged source image it doesn't need changed
// tighten up the top and left source image borders - this will enable
// border replication on the left and top borders of the original source
// image
oSrcOffset.x = 0;
oSrcOffset.y = 0;
// tighten up the right and bottom side source image borders - this will
// enable border replication on the right and bottom borders of the original
// source image
oEnlargedSrcSize.width = oSrcSize.width;
oEnlargedSrcSize.height = oSrcSize.height;
// create device output images for the mixed edge results
npp::ImageNPP_8u_C1 oDeviceDstOutXMixedBorders(oSizeROI.width,
oSizeROI.height);
npp::ImageNPP_8u_C1 oDeviceDstOutYMixedBorders(oSizeROI.width,
oSizeROI.height);
// shrink output ROI width so that only the left half of the destination
// image will be generated however since oEnlargedSrcSize.width is still set
// to oSrcSize.width then border control will be disabled when the filter
// needs to access source pixels beyond the right side of the left half of
// the source image
int nLeftWidth = oSizeROI.width / 2;
int nRightWidth = oSizeROI.width - nLeftWidth;
oSizeROI.width = nLeftWidth;
// run Prewitt edge detection gradient vector filter to generate the left
// side of the output image
NPP_CHECK_NPP(nppiGradientVectorPrewittBorder_8u16s_C1R(
pAdjustedSrc, oEnlargedDeviceSrc.pitch(), oEnlargedSrcSize, oSrcOffset,
oDeviceDstX.data(), oDeviceDstX.pitch(), oDeviceDstY.data(),
oDeviceDstY.pitch(), 0, 0, 0, 0, oSizeROI, NPP_MASK_SIZE_3_X_3,
nppiNormL1, NPP_BORDER_REPLICATE));
// now move the enlarged source pointer to the horizontal middle of the
// enlarged source image and tell the function where it was moved to
pAdjustedSrc += nLeftWidth;
// and adjust the source offset parameter accordingly - this will in effect
// turn off border control for the left border allowing the necessary source
// pixels to be used
oSrcOffset.x += nLeftWidth;
// update oSizeROI.width so that only enough destination pixels will be
// produced to fill the right half of the destination image
oSizeROI.width = nRightWidth;
// run Prewitt edge detection gradient vector filter to generate the right
// side of the output image adjusting the destination image pointers
// appropriately
NPP_CHECK_NPP(nppiGradientVectorPrewittBorder_8u16s_C1R(
pAdjustedSrc, oEnlargedDeviceSrc.pitch(), oEnlargedSrcSize, oSrcOffset,
oDeviceDstX.data() + nLeftWidth, oDeviceDstX.pitch(),
oDeviceDstY.data() + nLeftWidth, oDeviceDstY.pitch(), 0, 0, 0, 0,
oSizeROI, NPP_MASK_SIZE_3_X_3, nppiNormL1, NPP_BORDER_REPLICATE));
// convert 16s_C1 result images to binary 8u_C1 output images using constant
// value to adjust amount of visible detail
NPP_CHECK_NPP(nppiCompareC_16s_C1R(oDeviceDstX.data(), oDeviceDstX.pitch(),
32, oDeviceDstOutXMixedBorders.data(),
oDeviceDstOutXMixedBorders.pitch(),
oSizeROI, NPP_CMP_GREATER_EQ));
NPP_CHECK_NPP(nppiCompareC_16s_C1R(oDeviceDstY.data(), oDeviceDstY.pitch(),
32, oDeviceDstOutYMixedBorders.data(),
oDeviceDstOutYMixedBorders.pitch(),
oSizeROI, NPP_CMP_GREATER_EQ));
// create additional output files
std::string sResultXMixedBordersFilename = sResultBaseFilename;
std::string sResultYMixedBordersFilename = sResultBaseFilename;
sResultXMixedBordersFilename +=
"_gradientVectorPrewittBorderX_Vertical_WithMixedBorders.pgm";
sResultYMixedBordersFilename +=
"_gradientVectorPrewittBorderY_Horizontal_WithMixedBorders.pgm";
// copy the device result data into the host output images
oDeviceDstOutXMixedBorders.copyTo(oHostDstX.data(), oHostDstX.pitch());
oDeviceDstOutYMixedBorders.copyTo(oHostDstY.data(), oHostDstY.pitch());
saveImage(sResultXMixedBordersFilename, oHostDstX);
std::cout << "Saved image: " << sResultXMixedBordersFilename << std::endl;
saveImage(sResultYMixedBordersFilename, oHostDstY);
std::cout << "Saved image: " << sResultYMixedBordersFilename << std::endl;
// diff the original 8u_C1 result images with border control and the mixed
// border control images, they should match (diff image will be all black)
NPP_CHECK_NPP(nppiAbsDiff_8u_C1R(
oDeviceDstOutXMixedBorders.data(), oDeviceDstOutXMixedBorders.pitch(),
oDeviceDstOutX.data(), oDeviceDstOutX.pitch(),
oDeviceDstOutXDiff.data(), oDeviceDstOutXDiff.pitch(), oSizeROI));
NPP_CHECK_NPP(nppiAbsDiff_8u_C1R(
oDeviceDstOutYMixedBorders.data(), oDeviceDstOutYMixedBorders.pitch(),
oDeviceDstOutY.data(), oDeviceDstOutY.pitch(),
oDeviceDstOutYDiff.data(), oDeviceDstOutYDiff.pitch(), oSizeROI));
// create additional output files
std::string sResultXMixedDiffFilename = sResultBaseFilename;
std::string sResultYMixedDiffFilename = sResultBaseFilename;
sResultXMixedDiffFilename +=
"_gradientVectorPrewittBorderX_Vertical_MixedBorderDiffs.pgm";
sResultYMixedDiffFilename +=
"_gradientVectorPrewittBorderY_Horizontal_MixedBorderDiffs.pgm";
// copy the device result data into the host output images
oDeviceDstOutXDiff.copyTo(oHostDstX.data(), oHostDstX.pitch());
oDeviceDstOutYDiff.copyTo(oHostDstY.data(), oHostDstY.pitch());
saveImage(sResultXMixedDiffFilename, oHostDstX);
std::cout << "Saved image: " << sResultXMixedDiffFilename << std::endl;
saveImage(sResultYMixedDiffFilename, oHostDstY);
std::cout << "Saved image: " << sResultYMixedDiffFilename << std::endl;
nppiFree(oDeviceSrc.data());
nppiFree(oDeviceDstX.data());
nppiFree(oDeviceDstY.data());
nppiFree(oDeviceDstOutX.data());
nppiFree(oDeviceDstOutY.data());
nppiFree(oDeviceDstOutXNoBorders.data());
nppiFree(oDeviceDstOutYNoBorders.data());
nppiFree(oDeviceDstOutXDiff.data());
nppiFree(oDeviceDstOutYDiff.data());
nppiFree(oDeviceDstOutXMixedBorders.data());
nppiFree(oDeviceDstOutYMixedBorders.data());
nppiFree(oEnlargedDeviceSrc.data());
cudaDeviceReset();
exit(EXIT_SUCCESS);
} catch (npp::Exception &rException) {
std::cerr << "Program error! The following exception occurred: \n";
std::cerr << rException << std::endl;
std::cerr << "Aborting." << std::endl;
cudaDeviceReset();
exit(EXIT_FAILURE);
} catch (...) {
std::cerr << "Program error! An unknow type of exception occurred. \n";
std::cerr << "Aborting." << std::endl;
cudaDeviceReset();
exit(EXIT_FAILURE);
return -1;
}
return 0;
}

20
Samples/FilterBorderControlNPP/FilterBorderControlNPP_vs2017.sln Normal file
View File

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

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2017
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "FilterBorderControlNPP", "FilterBorderControlNPP_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

123
Samples/FilterBorderControlNPP/FilterBorderControlNPP_vs2017.vcxproj Normal file
View File

@@ -0,0 +1,123 @@
<?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>FilterBorderControlNPP_vs2017</RootNamespace>
<ProjectName>FilterBorderControlNPP</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 11.5.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;$(CudaToolkitIncludeDir);../../Common/UtilNPP;../common/FreeImage/Dist/x64;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>nppisu.lib;nppif.lib;nppitc.lib;nppidei.lib;nppial.lib;nppc.lib;cudart_static.lib;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies)</AdditionalDependencies>
<AdditionalLibraryDirectories>$(CudaToolkitLibDir);../common/FreeImage/Dist/x64;</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/FilterBorderControlNPP.exe</OutputFile>
</Link>
<CudaCompile>
<CodeGeneration>compute_35,compute_35;</CodeGeneration>
<AdditionalOptions>-Xcompiler "/wd 4819" %(AdditionalOptions)</AdditionalOptions>
<Include>./;../../Common</Include>
<Defines>WIN32</Defines>
<AdditionalOptions>--threads 0</AdditionalOptions>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(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="'$(Configuration)'=='Release'">
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
</Link>
<CudaCompile>
<Runtime>MT</Runtime>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Platform)'=='Win32'">
<Link>
</Link>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Platform)'=='x64'">
<Link>
<AdditionalDependencies>freeimage.lib;%(AdditionalDependencies)</AdditionalDependencies>
</Link>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="FilterBorderControlNPP.cpp" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

20
Samples/FilterBorderControlNPP/FilterBorderControlNPP_vs2019.sln Normal file
View File

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

Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio 2019
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "FilterBorderControlNPP", "FilterBorderControlNPP_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

119
Samples/FilterBorderControlNPP/FilterBorderControlNPP_vs2019.vcxproj Normal file
View File

@@ -0,0 +1,119 @@
<?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>FilterBorderControlNPP_vs2019</RootNamespace>
<ProjectName>FilterBorderControlNPP</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="$(CUDAPropsPath)\CUDA 11.5.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(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);../../Common/UtilNPP;../common/FreeImage/Dist/x64;</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<AdditionalDependencies>nppisu.lib;nppif.lib;nppitc.lib;nppidei.lib;nppial.lib;nppc.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>$(CudaToolkitLibDir);../common/FreeImage/Dist/x64;</AdditionalLibraryDirectories>
<OutputFile>$(OutDir)/FilterBorderControlNPP.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>
<ItemDefinitionGroup Condition="'$(Platform)'=='Win32'">
<Link>
</Link>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Platform)'=='x64'">
<Link>
<AdditionalDependencies>freeimage.lib;%(AdditionalDependencies)</AdditionalDependencies>
</Link>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="FilterBorderControlNPP.cpp" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.5.targets" />
</ImportGroup>
</Project>

378
Samples/FilterBorderControlNPP/Makefile Normal file
View File

@@ -0,0 +1,378 @@
################################################################################
# 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 - FilterBorderControlNPP 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 :=
################################################################################
#Detect if installed version of GCC supports required C++11
ifeq ($(TARGET_OS),linux)
empty :=
space := $(empty) $(empty)
GCCVERSIONSTRING := $(shell expr `$(HOST_COMPILER) -dumpversion`)
#Create version number without "."
GCCVERSION := $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f1 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f2 -d.)
GCCVERSION += $(shell expr `echo $(GCCVERSIONSTRING)` | cut -f3 -d.)
# Make sure the version number has at least 3 decimals
GCCVERSION += 00
# Remove spaces from the version number
GCCVERSION := $(subst $(space),$(empty),$(GCCVERSION))
#$(warning $(GCCVERSION))
IS_MIN_VERSION := $(shell expr `echo $(GCCVERSION)` \>= 51000)
ifeq ($(IS_MIN_VERSION), 1)
$(info >>> GCC Version is greater or equal to 5.1.0 <<<)
else
$(info >>> Waiving build. Minimum GCC version required is 5.1.0<<<)
SAMPLE_ENABLED := 0
endif
endif
# Gencode arguments
SMS ?=
ifeq ($(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 --std=c++11
INCLUDES += -I../../Common/UtilNPP
LIBRARIES += -lnppisu_static -lnppif_static -lnppitc_static -lnppidei_static -lnppial_static -lnppc_static -lculibos -lfreeimage
# Attempt to compile a minimal application linked against FreeImage. If a.out exists, FreeImage is properly set up.
$(shell echo "#include \"FreeImage.h\"" > test.c; echo "int main() { return 0; }" >> test.c ; $(NVCC) $(ALL_CCFLAGS) $(INCLUDES) $(ALL_LDFLAGS) $(LIBRARIES) -l freeimage test.c)
FREEIMAGE := $(shell find a.out 2>/dev/null)
$(shell rm a.out test.c 2>/dev/null)
ifeq ("$(FREEIMAGE)","")
$(info >>> WARNING - FreeImage is not set up correctly. Please ensure FreeImage is set up correctly. <<<)
SAMPLE_ENABLED := 0
endif
ifeq ($(SAMPLE_ENABLED),0)
EXEC ?= @echo "[@]"
endif
################################################################################
# Target rules
all: build
build: FilterBorderControlNPP
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
FilterBorderControlNPP.o:FilterBorderControlNPP.cpp
$(EXEC) $(NVCC) $(INCLUDES) $(ALL_CCFLAGS) $(GENCODE_FLAGS) -o $@ -c $<
FilterBorderControlNPP: FilterBorderControlNPP.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
$(EXEC) ./FilterBorderControlNPP
clean:
rm -f FilterBorderControlNPP FilterBorderControlNPP.o
rm -rf ../../bin/$(TARGET_ARCH)/$(TARGET_OS)/$(BUILD_TYPE)/FilterBorderControlNPP
clobber: clean

85
Samples/FilterBorderControlNPP/NsightEclipse.xml Normal file
View File

@@ -0,0 +1,85 @@
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
<name>FilterBorderControlNPP</name>
<description><![CDATA[This sample demonstrates how any border version of an NPP filtering function can be used in the most common mode, with border control enabled. Mentioned functions can be used to duplicate the results of the equivalent non-border version of the NPP functions. They can be also used for enabling and disabling border control on various source image edges depending on what portion of the source image is being used as input.]]></description>
<devicecompilation>whole</devicecompilation>
<fallback_min_ptx>true</fallback_min_ptx>
<includepaths>
<path>../../Common/UtilNPP</path>
<path os="Windows">../common/FreeImage/Dist/x64</path>
<path>./</path>
<path>../</path>
<path>../../Common</path>
</includepaths>
<keyconcepts>
<concept level="basic">Performance Strategies</concept>
<concept level="basic">Image Processing</concept>
<concept level="basic">NPP Library</concept>
</keyconcepts>
<keywords>
<keyword>CUDA</keyword>
<keyword>NPP</keyword>
<keyword>Image Processing</keyword>
</keywords>
<libraries>
<library>nppisu_static</library>
<library>nppif_static</library>
<library>nppitc_static</library>
<library>nppidei_static</library>
<library>nppial_static</library>
<library>nppc_static</library>
<library>culibos</library>
<library>freeimage</library>
</libraries>
<librarypaths>
</librarypaths>
<nsight_eclipse>true</nsight_eclipse>
<primary_file>FilterBorderControlNPP.cpp</primary_file>
<required_dependencies>
<dependency>FreeImage</dependency>
<dependency>NPP</dependency>
</required_dependencies>
<scopes>
<scope>1:CUDA Basic Topics</scope>
<scope>1:Performance Strategies</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>arm</arch>
</env>
<env>
<arch>ppc64le</arch>
<platform>linux</platform>
</env>
</supported_envs>
<supported_sm_architectures>
<include>all</include>
</supported_sm_architectures>
<title>Filter Border Control NPP</title>
<type>exe</type>
</entry>

71
Samples/FilterBorderControlNPP/README.md Normal file
View File

@@ -0,0 +1,71 @@
# FilterBorderControlNPP - Filter Border Control NPP
## Description
This sample demonstrates how any border version of an NPP filtering function can be used in the most common mode, with border control enabled. Mentioned functions can be used to duplicate the results of the equivalent non-border version of the NPP functions. They can be also used for enabling and disabling border control on various source image edges depending on what portion of the source image is being used as input.
## Key Concepts
Performance Strategies, Image Processing, NPP 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
[FreeImage](../../README.md#freeimage), [NPP](../../README.md#npp)
## Prerequisites
Download and install the [CUDA Toolkit 11.5](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)