NASP/cuda-samples
83
0
Fork 0
mirror of https://github.com/NVIDIA/cuda-samples.git synced 2026-10-11 23:38:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Initial public release for CUDA 9.2.

Browse Source
This commit is contained in:
Andy Dick
2018-03-02 16:07:37 -08:00
commit 8bb8c5fac0
111 changed files with 14869 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

470
Common/drvapi_error_string.h Normal file
View File

@@ -0,0 +1,470 @@
/* Copyright (c) 2018, 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 COMMON_DRVAPI_ERROR_STRING_H_
#define COMMON_DRVAPI_ERROR_STRING_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#ifdef __cuda_cuda_h__ // check to see if CUDA_H is included above
// Error Code string definitions here
typedef struct {
char const *error_string;
int error_id;
} s_CudaErrorStr;
/**
* Error codes
*/
static s_CudaErrorStr sCudaDrvErrorString[] = {
/**
* The API call returned with no errors. In the case of query calls, this
* can also mean that the operation being queried is complete (see
* ::cuEventQuery() and ::cuStreamQuery()).
*/
{"CUDA_SUCCESS", 0},
/**
* This indicates that one or more of the parameters passed to the API call
* is not within an acceptable range of values.
*/
{"CUDA_ERROR_INVALID_VALUE", 1},
/**
* The API call failed because it was unable to allocate enough memory to
* perform the requested operation.
*/
{"CUDA_ERROR_OUT_OF_MEMORY", 2},
/**
* This indicates that the CUDA driver has not been initialized with
* ::cuInit() or that initialization has failed.
*/
{"CUDA_ERROR_NOT_INITIALIZED", 3},
/**
* This indicates that the CUDA driver is in the process of shutting down.
*/
{"CUDA_ERROR_DEINITIALIZED", 4},
/**
* This indicates profiling APIs are called while application is running
* in visual profiler mode.
*/
{"CUDA_ERROR_PROFILER_DISABLED", 5},
/**
* This indicates profiling has not been initialized for this context.
* Call cuProfilerInitialize() to resolve this.
*/
{"CUDA_ERROR_PROFILER_NOT_INITIALIZED", 6},
/**
* This indicates profiler has already been started and probably
* cuProfilerStart() is incorrectly called.
*/
{"CUDA_ERROR_PROFILER_ALREADY_STARTED", 7},
/**
* This indicates profiler has already been stopped and probably
* cuProfilerStop() is incorrectly called.
*/
{"CUDA_ERROR_PROFILER_ALREADY_STOPPED", 8},
/**
* This indicates that no CUDA-capable devices were detected by the
* installed CUDA driver.
*/
{"CUDA_ERROR_NO_DEVICE (no CUDA-capable devices were detected)", 100},
/**
* This indicates that the device ordinal supplied by the user does not
* correspond to a valid CUDA device.
*/
{"CUDA_ERROR_INVALID_DEVICE (device specified is not a valid CUDA device)",
101},
/**
* This indicates that the device kernel image is invalid. This can also
* indicate an invalid CUDA module.
*/
{"CUDA_ERROR_INVALID_IMAGE", 200},
/**
* This most frequently indicates that there is no context bound to the
* current thread. This can also be returned if the context passed to an
* API call is not a valid handle (such as a context that has had
* ::cuCtxDestroy() invoked on it). This can also be returned if a user
* mixes different API versions (i.e. 3010 context with 3020 API calls).
* See ::cuCtxGetApiVersion() for more details.
*/
{"CUDA_ERROR_INVALID_CONTEXT", 201},
/**
* This indicated that the context being supplied as a parameter to the
* API call was already the active context.
* \deprecated
* This error return is deprecated as of CUDA 3.2. It is no longer an
* error to attempt to push the active context via ::cuCtxPushCurrent().
*/
{"CUDA_ERROR_CONTEXT_ALREADY_CURRENT", 202},
/**
* This indicates that a map or register operation has failed.
*/
{"CUDA_ERROR_MAP_FAILED", 205},
/**
* This indicates that an unmap or unregister operation has failed.
*/
{"CUDA_ERROR_UNMAP_FAILED", 206},
/**
* This indicates that the specified array is currently mapped and thus
* cannot be destroyed.
*/
{"CUDA_ERROR_ARRAY_IS_MAPPED", 207},
/**
* This indicates that the resource is already mapped.
*/
{"CUDA_ERROR_ALREADY_MAPPED", 208},
/**
* This indicates that there is no kernel image available that is suitable
* for the device. This can occur when a user specifies code generation
* options for a particular CUDA source file that do not include the
* corresponding device configuration.
*/
{"CUDA_ERROR_NO_BINARY_FOR_GPU", 209},
/**
* This indicates that a resource has already been acquired.
*/
{"CUDA_ERROR_ALREADY_ACQUIRED", 210},
/**
* This indicates that a resource is not mapped.
*/
{"CUDA_ERROR_NOT_MAPPED", 211},
/**
* This indicates that a mapped resource is not available for access as an
* array.
*/
{"CUDA_ERROR_NOT_MAPPED_AS_ARRAY", 212},
/**
* This indicates that a mapped resource is not available for access as a
* pointer.
*/
{"CUDA_ERROR_NOT_MAPPED_AS_POINTER", 213},
/**
* This indicates that an uncorrectable ECC error was detected during
* execution.
*/
{"CUDA_ERROR_ECC_UNCORRECTABLE", 214},
/**
* This indicates that the ::CUlimit passed to the API call is not
* supported by the active device.
*/
{"CUDA_ERROR_UNSUPPORTED_LIMIT", 215},
/**
* This indicates that the ::CUcontext passed to the API call can
* only be bound to a single CPU thread at a time but is already
* bound to a CPU thread.
*/
{"CUDA_ERROR_CONTEXT_ALREADY_IN_USE", 216},
/**
* This indicates that peer access is not supported across the given
* devices.
*/
{"CUDA_ERROR_PEER_ACCESS_UNSUPPORTED", 217},
/**
* This indicates that a PTX JIT compilation failed.
*/
{"CUDA_ERROR_INVALID_PTX", 218},
/**
* This indicates an error with OpenGL or DirectX context.
*/
{"CUDA_ERROR_INVALID_GRAPHICS_CONTEXT", 219},
/**
* This indicates that an uncorrectable NVLink error was detected during the
* execution.
*/
{"CUDA_ERROR_NVLINK_UNCORRECTABLE", 220},
/**
* This indicates that the PTX JIT compiler library was not found.
*/
{"CUDA_ERROR_JIT_COMPILER_NOT_FOUND", 221},
/**
* This indicates that the device kernel source is invalid.
*/
{"CUDA_ERROR_INVALID_SOURCE", 300},
/**
* This indicates that the file specified was not found.
*/
{"CUDA_ERROR_FILE_NOT_FOUND", 301},
/**
* This indicates that a link to a shared object failed to resolve.
*/
{"CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND", 302},
/**
* This indicates that initialization of a shared object failed.
*/
{"CUDA_ERROR_SHARED_OBJECT_INIT_FAILED", 303},
/**
* This indicates that an OS call failed.
*/
{"CUDA_ERROR_OPERATING_SYSTEM", 304},
/**
* This indicates that a resource handle passed to the API call was not
* valid. Resource handles are opaque types like ::CUstream and ::CUevent.
*/
{"CUDA_ERROR_INVALID_HANDLE", 400},
/**
* This indicates that a named symbol was not found. Examples of symbols
* are global/constant variable names, texture names }, and surface names.
*/
{"CUDA_ERROR_NOT_FOUND", 500},
/**
* This indicates that asynchronous operations issued previously have not
* completed yet. This result is not actually an error, but must be
* indicated differently than ::CUDA_SUCCESS (which indicates completion).
* Calls that may return this value include ::cuEventQuery() and
* ::cuStreamQuery().
*/
{"CUDA_ERROR_NOT_READY", 600},
/**
* While executing a kernel, the device encountered a
* load or store instruction on an invalid memory address.
* This leaves the process in an inconsistent state and any further CUDA
* work will return the same error. To continue using CUDA, the process must
* be terminated and relaunched.
*/
{"CUDA_ERROR_ILLEGAL_ADDRESS", 700},
/**
* This indicates that a launch did not occur because it did not have
* appropriate resources. This error usually indicates that the user has
* attempted to pass too many arguments to the device kernel, or the
* kernel launch specifies too many threads for the kernel's register
* count. Passing arguments of the wrong size (i.e. a 64-bit pointer
* when a 32-bit int is expected) is equivalent to passing too many
* arguments and can also result in this error.
*/
{"CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES", 701},
/**
* This indicates that the device kernel took too long to execute. This can
* only occur if timeouts are enabled - see the device attribute
* ::CU_DEVICE_ATTRIBUTE_KERNEL_EXEC_TIMEOUT for more information. The
* context cannot be used (and must be destroyed similar to
* ::CUDA_ERROR_LAUNCH_FAILED). All existing device memory allocations from
* this context are invalid and must be reconstructed if the program is to
* continue using CUDA.
*/
{"CUDA_ERROR_LAUNCH_TIMEOUT", 702},
/**
* This error indicates a kernel launch that uses an incompatible texturing
* mode.
*/
{"CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING", 703},
/**
* This error indicates that a call to ::cuCtxEnablePeerAccess() is
* trying to re-enable peer access to a context which has already
* had peer access to it enabled.
*/
{"CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED", 704},
/**
* This error indicates that ::cuCtxDisablePeerAccess() is
* trying to disable peer access which has not been enabled yet
* via ::cuCtxEnablePeerAccess().
*/
{"CUDA_ERROR_PEER_ACCESS_NOT_ENABLED", 705},
/**
* This error indicates that the primary context for the specified device
* has already been initialized.
*/
{"CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE", 708},
/**
* This error indicates that the context current to the calling thread
* has been destroyed using ::cuCtxDestroy }, or is a primary context which
* has not yet been initialized.
*/
{"CUDA_ERROR_CONTEXT_IS_DESTROYED", 709},
/**
* A device-side assert triggered during kernel execution. The context
* cannot be used anymore, and must be destroyed. All existing device
* memory allocations from this context are invalid and must be
* reconstructed if the program is to continue using CUDA.
*/
{"CUDA_ERROR_ASSERT", 710},
/**
* This error indicates that the hardware resources required to enable
* peer access have been exhausted for one or more of the devices
* passed to ::cuCtxEnablePeerAccess().
*/
{"CUDA_ERROR_TOO_MANY_PEERS", 711},
/**
* This error indicates that the memory range passed to
* ::cuMemHostRegister() has already been registered.
*/
{"CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED", 712},
/**
* This error indicates that the pointer passed to ::cuMemHostUnregister()
* does not correspond to any currently registered memory region.
*/
{"CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED", 713},
/**
* While executing a kernel, the device encountered a stack error.
* This can be due to stack corruption or exceeding the stack size limit.
* This leaves the process in an inconsistent state and any further CUDA
* work will return the same error. To continue using CUDA, the process must
* be terminated and relaunched.
*/
{"CUDA_ERROR_HARDWARE_STACK_ERROR", 714},
/**
* While executing a kernel, the device encountered an illegal instruction.
* This leaves the process in an inconsistent state and any further CUDA
* work will return the same error. To continue using CUDA, the process must
* be terminated and relaunched.
*/
{"CUDA_ERROR_ILLEGAL_INSTRUCTION", 715},
/**
* While executing a kernel, the device encountered a load or store
* instruction on a memory address which is not aligned. This leaves the
* process in an inconsistent state and any further CUDA work will return
* the same error. To continue using CUDA, the process must be terminated
* and relaunched.
*/
{"CUDA_ERROR_MISALIGNED_ADDRESS", 716},
/**
* While executing a kernel, the device encountered an instruction
* which can only operate on memory locations in certain address spaces
* (global, shared, or local), but was supplied a memory address not
* belonging to an allowed address space.
* This leaves the process in an inconsistent state and any further CUDA
* work will return the same error. To continue using CUDA, the process must
* be terminated and relaunched.
*/
{"CUDA_ERROR_INVALID_ADDRESS_SPACE", 717},
/**
* While executing a kernel, the device program counter wrapped its address
* space. This leaves the process in an inconsistent state and any further
* CUDA work will return the same error. To continue using CUDA, the process
* must be terminated and relaunched.
*/
{"CUDA_ERROR_INVALID_PC", 718},
/**
* An exception occurred on the device while executing a kernel. Common
* causes include dereferencing an invalid device pointer and accessing
* out of bounds shared memory. The context cannot be used }, so it must
* be destroyed (and a new one should be created). All existing device
* memory allocations from this context are invalid and must be
* reconstructed if the program is to continue using CUDA.
*/
{"CUDA_ERROR_LAUNCH_FAILED", 719},
/**
* This error indicates that the number of blocks launched per grid for a
* kernel that was launched via either ::cuLaunchCooperativeKernel or
* ::cuLaunchCooperativeKernelMultiDevice exceeds the maximum number of
* blocks as allowed by ::cuOccupancyMaxActiveBlocksPerMultiprocessor or
* ::cuOccupancyMaxActiveBlocksPerMultiprocessorWithFlags times the number
* of multiprocessors as specified by the device attribute
* ::CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT.
*/
{"CUDA_ERROR_COOPERATIVE_LAUNCH_TOO_LARGE", 720},
/**
* This error indicates that the attempted operation is not permitted.
*/
{"CUDA_ERROR_NOT_PERMITTED", 800},
/**
* This error indicates that the attempted operation is not supported
* on the current system or device.
*/
{"CUDA_ERROR_NOT_SUPPORTED", 801},
/**
* This indicates that an unknown internal error has occurred.
*/
{"CUDA_ERROR_UNKNOWN", 999},
{NULL, -1}};
// This is just a linear search through the array, since the error_id's are not
// always ocurring consecutively
inline const char *getCudaDrvErrorString(CUresult error_id) {
int index = 0;
while (sCudaDrvErrorString[index].error_id != error_id &&
sCudaDrvErrorString[index].error_id != -1) {
index++;
}
if (sCudaDrvErrorString[index].error_id == error_id)
return (const char *)sCudaDrvErrorString[index].error_string;
else
return (const char *)"CUDA_ERROR not found!";
}
#endif // __cuda_cuda_h__
#endif // COMMON_DRVAPI_ERROR_STRING_H_

151
Common/exception.h Normal file
View File

@@ -0,0 +1,151 @@
/* Copyright (c) 2018, 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.
*/
/* CUda UTility Library */
#ifndef COMMON_EXCEPTION_H_
#define COMMON_EXCEPTION_H_
// includes, system
#include <stdlib.h>
#include <exception>
#include <iostream>
#include <stdexcept>
#include <string>
//! Exception wrapper.
//! @param Std_Exception Exception out of namespace std for easy typing.
template <class Std_Exception>
class Exception : public Std_Exception {
public:
//! @brief Static construction interface
//! @return Alwayss throws ( Located_Exception<Exception>)
//! @param file file in which the Exception occurs
//! @param line line in which the Exception occurs
//! @param detailed details on the code fragment causing the Exception
static void throw_it(const char *file, const int line,
const char *detailed = "-");
//! Static construction interface
//! @return Alwayss throws ( Located_Exception<Exception>)
//! @param file file in which the Exception occurs
//! @param line line in which the Exception occurs
//! @param detailed details on the code fragment causing the Exception
static void throw_it(const char *file, const int line,
const std::string &detailed);
//! Destructor
virtual ~Exception() throw();
private:
//! Constructor, default (private)
Exception();
//! Constructor, standard
//! @param str string returned by what()
explicit Exception(const std::string &str);
};
////////////////////////////////////////////////////////////////////////////////
//! Exception handler function for arbitrary exceptions
//! @param ex exception to handle
////////////////////////////////////////////////////////////////////////////////
template <class Exception_Typ>
inline void handleException(const Exception_Typ &ex) {
std::cerr << ex.what() << std::endl;
exit(EXIT_FAILURE);
}
//! Convenience macros
//! Exception caused by dynamic program behavior, e.g. file does not exist
#define RUNTIME_EXCEPTION(msg) \
Exception<std::runtime_error>::throw_it(__FILE__, __LINE__, msg)
//! Logic exception in program, e.g. an assert failed
#define LOGIC_EXCEPTION(msg) \
Exception<std::logic_error>::throw_it(__FILE__, __LINE__, msg)
//! Out of range exception
#define RANGE_EXCEPTION(msg) \
Exception<std::range_error>::throw_it(__FILE__, __LINE__, msg)
////////////////////////////////////////////////////////////////////////////////
//! Implementation
// includes, system
#include <sstream>
////////////////////////////////////////////////////////////////////////////////
//! Static construction interface.
//! @param Exception causing code fragment (file and line) and detailed infos.
////////////////////////////////////////////////////////////////////////////////
/*static*/ template <class Std_Exception>
void Exception<Std_Exception>::throw_it(const char *file, const int line,
const char *detailed) {
std::stringstream s;
// Quiet heavy-weight but exceptions are not for
// performance / release versions
s << "Exception in file '" << file << "' in line " << line << "\n"
<< "Detailed description: " << detailed << "\n";
throw Exception(s.str());
}
////////////////////////////////////////////////////////////////////////////////
//! Static construction interface.
//! @param Exception causing code fragment (file and line) and detailed infos.
////////////////////////////////////////////////////////////////////////////////
/*static*/ template <class Std_Exception>
void Exception<Std_Exception>::throw_it(const char *file, const int line,
const std::string &msg) {
throw_it(file, line, msg.c_str());
}
////////////////////////////////////////////////////////////////////////////////
//! Constructor, default (private).
////////////////////////////////////////////////////////////////////////////////
template <class Std_Exception>
Exception<Std_Exception>::Exception() : Std_Exception("Unknown Exception.\n") {}
////////////////////////////////////////////////////////////////////////////////
//! Constructor, standard (private).
//! String returned by what().
////////////////////////////////////////////////////////////////////////////////
template <class Std_Exception>
Exception<Std_Exception>::Exception(const std::string &s) : Std_Exception(s) {}
////////////////////////////////////////////////////////////////////////////////
//! Destructor
////////////////////////////////////////////////////////////////////////////////
template <class Std_Exception>
Exception<Std_Exception>::~Exception() throw() {}
// functions, exported
#endif // COMMON_EXCEPTION_H_

1341
Common/helper_cuda.h Normal file
View File

File diff suppressed because it is too large Load Diff

419
Common/helper_cuda_drvapi.h Normal file
View File

@@ -0,0 +1,419 @@
/* Copyright (c) 2018, 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.
*/
// Helper functions for CUDA Driver API error handling (make sure that CUDA_H is
// included in your projects)
#ifndef COMMON_HELPER_CUDA_DRVAPI_H_
#define COMMON_HELPER_CUDA_DRVAPI_H_
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <drvapi_error_string.h>
#include <helper_string.h>
#ifndef MAX
#define MAX(a, b) (a > b ? a : b)
#endif
#ifndef COMMON_HELPER_CUDA_H_
inline int ftoi(float value) {
return (value >= 0 ? static_cast<int>(value + 0.5)
: static_cast<int>(value - 0.5));
}
#endif
#ifndef EXIT_WAIVED
#define EXIT_WAIVED 2
#endif
////////////////////////////////////////////////////////////////////////////////
// These are CUDA Helper functions
// add a level of protection to the CUDA SDK samples, let's force samples to
// explicitly include CUDA.H
#ifdef __cuda_cuda_h__
// This will output the proper CUDA error strings in the event that a CUDA host
// call returns an error
#ifndef checkCudaErrors
#define checkCudaErrors(err) __checkCudaErrors(err, __FILE__, __LINE__)
// These are the inline versions for all of the SDK helper functions
inline void __checkCudaErrors(CUresult err, const char *file, const int line) {
if (CUDA_SUCCESS != err) {
fprintf(stderr,
"checkCudaErrors() Driver API error = %04d \"%s\" from file <%s>, "
"line %i.\n",
err, getCudaDrvErrorString(err), file, line);
exit(EXIT_FAILURE);
}
}
#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
template <class T>
inline void getCudaAttribute(T *attribute, CUdevice_attribute device_attribute,
int device) {
CUresult error_result =
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
// Beginning of GPU Architecture definitions
inline int _ConvertSMVer2CoresDRV(int major, int minor) {
// Defines for GPU Architecture types (using the SM version to determine the #
// of cores per SM
typedef struct {
int SM; // 0xMm (hexidecimal notation), M = SM Major version, and m = SM
// minor version
int Cores;
} sSMtoCores;
sSMtoCores nGpuArchCoresPerSM[] = {
{0x30, 192}, // Kepler Generation (SM 3.0) GK10x class
{0x32, 192}, // Kepler Generation (SM 3.2) GK10x class
{0x35, 192}, // Kepler Generation (SM 3.5) GK11x class
{0x37, 192}, // Kepler Generation (SM 3.7) GK21x class
{0x50, 128}, // Maxwell Generation (SM 5.0) GM10x class
{0x52, 128}, // Maxwell Generation (SM 5.2) GM20x class
{0x53, 128}, // Maxwell Generation (SM 5.3) GM20x class
{0x60, 64}, // Pascal Generation (SM 6.0) GP100 class
{0x61, 128}, // Pascal Generation (SM 6.1) GP10x class
{0x62, 128}, // Pascal Generation (SM 6.2) GP10x class
{0x70, 64}, // Volta Generation (SM 7.0) GV100 class
{0x72, 64}, // Volta Generation (SM 7.2) GV11b class
{-1, -1}};
int index = 0;
while (nGpuArchCoresPerSM[index].SM != -1) {
if (nGpuArchCoresPerSM[index].SM == ((major << 4) + minor)) {
return nGpuArchCoresPerSM[index].Cores;
}
index++;
}
// If we don't find the values, we default use the previous one to run
// properly
printf(
"MapSMtoCores for SM %d.%d is undefined. Default to use %d Cores/SM\n",
major, minor, nGpuArchCoresPerSM[index - 1].Cores);
return nGpuArchCoresPerSM[index - 1].Cores;
}
// end of GPU Architecture definitions
#ifdef __cuda_cuda_h__
// General GPU Device CUDA Initialization
inline int gpuDeviceInitDRV(int ARGC, const char **ARGV) {
int cuDevice = 0;
int deviceCount = 0;
CUresult err = cuInit(0);
if (CUDA_SUCCESS == err) {
checkCudaErrors(cuDeviceGetCount(&deviceCount));
}
if (deviceCount == 0) {
fprintf(stderr, "cudaDeviceInit error: no devices supporting CUDA\n");
exit(EXIT_FAILURE);
}
int dev = 0;
dev = getCmdLineArgumentInt(ARGC, (const char **)ARGV, "device=");
if (dev < 0) {
dev = 0;
}
if (dev > deviceCount - 1) {
fprintf(stderr, "\n");
fprintf(stderr, ">> %d CUDA capable GPU device(s) detected. <<\n",
deviceCount);
fprintf(stderr,
">> cudaDeviceInit (-device=%d) is not a valid GPU device. <<\n",
dev);
fprintf(stderr, "\n");
return -dev;
}
checkCudaErrors(cuDeviceGet(&cuDevice, dev));
char name[100];
cuDeviceGetName(name, 100, cuDevice);
int computeMode;
getCudaAttribute<int>(&computeMode, CU_DEVICE_ATTRIBUTE_COMPUTE_MODE, dev);
if (computeMode == CU_COMPUTEMODE_PROHIBITED) {
fprintf(stderr,
"Error: device is running in <CU_COMPUTEMODE_PROHIBITED>, no "
"threads can use this CUDA Device.\n");
return -1;
}
if (checkCmdLineFlag(ARGC, (const char **)ARGV, "quiet") == false) {
printf("gpuDeviceInitDRV() Using CUDA Device [%d]: %s\n", dev, name);
}
return dev;
}
// This function returns the best GPU based on performance
inline int gpuGetMaxGflopsDeviceIdDRV() {
CUdevice current_device = 0;
CUdevice max_perf_device = 0;
int device_count = 0;
int sm_per_multiproc = 0;
unsigned long long max_compute_perf = 0;
int best_SM_arch = 0;
int major = 0;
int minor = 0;
int multiProcessorCount;
int clockRate;
int devices_prohibited = 0;
cuInit(0);
checkCudaErrors(cuDeviceGetCount(&device_count));
if (device_count == 0) {
fprintf(stderr,
"gpuGetMaxGflopsDeviceIdDRV error: no devices supporting CUDA\n");
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
current_device = 0;
while (current_device < device_count) {
checkCudaErrors(cuDeviceGetAttribute(
&multiProcessorCount, CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT,
current_device));
checkCudaErrors(cuDeviceGetAttribute(
&clockRate, CU_DEVICE_ATTRIBUTE_CLOCK_RATE, current_device));
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, current_device));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, current_device));
int computeMode;
getCudaAttribute<int>(&computeMode, CU_DEVICE_ATTRIBUTE_COMPUTE_MODE,
current_device);
if (computeMode != CU_COMPUTEMODE_PROHIBITED) {
if (major == 9999 && minor == 9999) {
sm_per_multiproc = 1;
} else {
sm_per_multiproc = _ConvertSMVer2CoresDRV(major, minor);
}
unsigned long long compute_perf =
(unsigned long long)(multiProcessorCount * sm_per_multiproc *
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 (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;
}
}
} else {
devices_prohibited++;
}
++current_device;
}
if (devices_prohibited == device_count) {
fprintf(stderr,
"gpuGetMaxGflopsDeviceIdDRV error: all devices have compute mode "
"prohibited.\n");
exit(EXIT_FAILURE);
}
return max_perf_device;
}
// General initialization call to pick the best CUDA Device
inline CUdevice findCudaDeviceDRV(int argc, const char **argv) {
CUdevice cuDevice;
int devID = 0;
// If the command-line has a device number specified, use it
if (checkCmdLineFlag(argc, (const char **)argv, "device")) {
devID = gpuDeviceInitDRV(argc, argv);
if (devID < 0) {
printf("exiting...\n");
exit(EXIT_SUCCESS);
}
} else {
// Otherwise pick the device with highest Gflops/s
char name[100];
devID = gpuGetMaxGflopsDeviceIdDRV();
checkCudaErrors(cuDeviceGet(&cuDevice, devID));
cuDeviceGetName(name, 100, cuDevice);
printf("> Using CUDA Device [%d]: %s\n", devID, name);
}
cuDeviceGet(&cuDevice, devID);
return cuDevice;
}
inline CUdevice findIntegratedGPUDrv() {
CUdevice current_device = 0;
int device_count = 0;
int devices_prohibited = 0;
int isIntegrated;
cuInit(0);
checkCudaErrors(cuDeviceGetCount(&device_count));
if (device_count == 0) {
fprintf(stderr, "CUDA error: no devices supporting CUDA.\n");
exit(EXIT_FAILURE);
}
// Find the integrated GPU which is compute capable
while (current_device < device_count) {
int computeMode = -1;
checkCudaErrors(cuDeviceGetAttribute(
&isIntegrated, CU_DEVICE_ATTRIBUTE_INTEGRATED, current_device));
checkCudaErrors(cuDeviceGetAttribute(
&computeMode, CU_DEVICE_ATTRIBUTE_COMPUTE_MODE, current_device));
// If GPU is integrated and is not running on Compute Mode prohibited use
// that
if (isIntegrated && (computeMode != CU_COMPUTEMODE_PROHIBITED)) {
int major = 0, minor = 0;
char deviceName[256];
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR,
current_device));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR,
current_device));
checkCudaErrors(cuDeviceGetName(deviceName, 256, current_device));
printf("GPU Device %d: \"%s\" with compute capability %d.%d\n\n",
current_device, deviceName, major, minor);
return current_device;
} else {
devices_prohibited++;
}
current_device++;
}
if (devices_prohibited == device_count) {
fprintf(stderr, "CUDA error: No Integrated CUDA capable GPU found.\n");
exit(EXIT_FAILURE);
}
return -1;
}
// General check for CUDA GPU SM Capabilities
inline bool checkCudaCapabilitiesDRV(int major_version, int minor_version,
int devID) {
CUdevice cuDevice;
char name[256];
int major = 0, minor = 0;
checkCudaErrors(cuDeviceGet(&cuDevice, devID));
checkCudaErrors(cuDeviceGetName(name, 100, cuDevice));
checkCudaErrors(cuDeviceGetAttribute(
&major, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, cuDevice));
checkCudaErrors(cuDeviceGetAttribute(
&minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, cuDevice));
if ((major > major_version) ||
(major == major_version && minor >= minor_version)) {
printf("> Device %d: <%16s >, Compute SM %d.%d detected\n", devID, name,
major, minor);
return true;
} else {
printf(
"No GPU device was found that can support CUDA compute capability "
"%d.%d.\n",
major_version, minor_version);
return false;
}
}
#endif
// end of CUDA Helper Functions
#endif // COMMON_HELPER_CUDA_DRVAPI_H_

59
Common/helper_functions.h Normal file
View File

@@ -0,0 +1,59 @@
/* Copyright (c) 2018, 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 (string parsing,
// timers, image helpers, etc)
#ifndef COMMON_HELPER_FUNCTIONS_H_
#define COMMON_HELPER_FUNCTIONS_H_
#ifdef WIN32
#pragma warning(disable : 4996)
#endif
// includes, project
#include <assert.h>
#include <exception.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <algorithm>
#include <fstream>
#include <iostream>
#include <string>
#include <vector>
// includes, timer, string parsing, image helpers
#include <helper_image.h> // helper functions for image compare, dump, data comparisons
#include <helper_string.h> // helper functions for string parsing
#include <helper_timer.h> // helper functions for timers
#ifndef EXIT_WAIVED
#define EXIT_WAIVED 2
#endif
#endif // COMMON_HELPER_FUNCTIONS_H_

1001
Common/helper_image.h Normal file
View File

File diff suppressed because it is too large Load Diff

699
Common/helper_string.h Normal file
View File

@@ -0,0 +1,699 @@
/* Copyright (c) 2018, 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 (string parsing, timers, etc)
#ifndef COMMON_HELPER_STRING_H_
#define COMMON_HELPER_STRING_H_
#include <stdio.h>
#include <stdlib.h>
#include <fstream>
#include <string>
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
#ifndef _CRT_SECURE_NO_DEPRECATE
#define _CRT_SECURE_NO_DEPRECATE
#endif
#ifndef STRCASECMP
#define STRCASECMP _stricmp
#endif
#ifndef STRNCASECMP
#define STRNCASECMP _strnicmp
#endif
#ifndef STRCPY
#define STRCPY(sFilePath, nLength, sPath) strcpy_s(sFilePath, nLength, sPath)
#endif
#ifndef FOPEN
#define FOPEN(fHandle, filename, mode) fopen_s(&fHandle, filename, mode)
#endif
#ifndef FOPEN_FAIL
#define FOPEN_FAIL(result) (result != 0)
#endif
#ifndef SSCANF
#define SSCANF sscanf_s
#endif
#ifndef SPRINTF
#define SPRINTF sprintf_s
#endif
#else // Linux Includes
#include <string.h>
#include <strings.h>
#ifndef STRCASECMP
#define STRCASECMP strcasecmp
#endif
#ifndef STRNCASECMP
#define STRNCASECMP strncasecmp
#endif
#ifndef STRCPY
#define STRCPY(sFilePath, nLength, sPath) strcpy(sFilePath, sPath)
#endif
#ifndef FOPEN
#define FOPEN(fHandle, filename, mode) (fHandle = fopen(filename, mode))
#endif
#ifndef FOPEN_FAIL
#define FOPEN_FAIL(result) (result == NULL)
#endif
#ifndef SSCANF
#define SSCANF sscanf
#endif
#ifndef SPRINTF
#define SPRINTF sprintf
#endif
#endif
#ifndef EXIT_WAIVED
#define EXIT_WAIVED 2
#endif
// CUDA Utility Helper Functions
inline int stringRemoveDelimiter(char delimiter, const char *string) {
int string_start = 0;
while (string[string_start] == delimiter) {
string_start++;
}
if (string_start >= static_cast<int>(strlen(string) - 1)) {
return 0;
}
return string_start;
}
inline int getFileExtension(char *filename, char **extension) {
int string_length = static_cast<int>(strlen(filename));
while (filename[string_length--] != '.') {
if (string_length == 0) break;
}
if (string_length > 0) string_length += 2;
if (string_length == 0)
*extension = NULL;
else
*extension = &filename[string_length];
return string_length;
}
inline bool checkCmdLineFlag(const int argc, const char **argv,
const char *string_ref) {
bool bFound = false;
if (argc >= 1) {
for (int i = 1; i < argc; i++) {
int string_start = stringRemoveDelimiter('-', argv[i]);
const char *string_argv = &argv[i][string_start];
const char *equal_pos = strchr(string_argv, '=');
int argv_length = static_cast<int>(
equal_pos == 0 ? strlen(string_argv) : equal_pos - string_argv);
int length = static_cast<int>(strlen(string_ref));
if (length == argv_length &&
!STRNCASECMP(string_argv, string_ref, length)) {
bFound = true;
continue;
}
}
}
return bFound;
}
// This function wraps the CUDA Driver API into a template function
template <class T>
inline bool getCmdLineArgumentValue(const int argc, const char **argv,
const char *string_ref, T *value) {
bool bFound = false;
if (argc >= 1) {
for (int i = 1; i < argc; i++) {
int string_start = stringRemoveDelimiter('-', argv[i]);
const char *string_argv = &argv[i][string_start];
int length = static_cast<int>(strlen(string_ref));
if (!STRNCASECMP(string_argv, string_ref, length)) {
if (length + 1 <= static_cast<int>(strlen(string_argv))) {
int auto_inc = (string_argv[length] == '=') ? 1 : 0;
*value = (T)atoi(&string_argv[length + auto_inc]);
}
bFound = true;
i = argc;
}
}
}
return bFound;
}
inline int getCmdLineArgumentInt(const int argc, const char **argv,
const char *string_ref) {
bool bFound = false;
int value = -1;
if (argc >= 1) {
for (int i = 1; i < argc; i++) {
int string_start = stringRemoveDelimiter('-', argv[i]);
const char *string_argv = &argv[i][string_start];
int length = static_cast<int>(strlen(string_ref));
if (!STRNCASECMP(string_argv, string_ref, length)) {
if (length + 1 <= static_cast<int>(strlen(string_argv))) {
int auto_inc = (string_argv[length] == '=') ? 1 : 0;
value = atoi(&string_argv[length + auto_inc]);
} else {
value = 0;
}
bFound = true;
continue;
}
}
}
if (bFound) {
return value;
} else {
return 0;
}
}
inline float getCmdLineArgumentFloat(const int argc, const char **argv,
const char *string_ref) {
bool bFound = false;
float value = -1;
if (argc >= 1) {
for (int i = 1; i < argc; i++) {
int string_start = stringRemoveDelimiter('-', argv[i]);
const char *string_argv = &argv[i][string_start];
int length = static_cast<int>(strlen(string_ref));
if (!STRNCASECMP(string_argv, string_ref, length)) {
if (length + 1 <= static_cast<int>(strlen(string_argv))) {
int auto_inc = (string_argv[length] == '=') ? 1 : 0;
value = static_cast<float>(atof(&string_argv[length + auto_inc]));
} else {
value = 0.f;
}
bFound = true;
continue;
}
}
}
if (bFound) {
return value;
} else {
return 0;
}
}
inline bool getCmdLineArgumentString(const int argc, const char **argv,
const char *string_ref,
char **string_retval) {
bool bFound = false;
if (argc >= 1) {
for (int i = 1; i < argc; i++) {
int string_start = stringRemoveDelimiter('-', argv[i]);
char *string_argv = const_cast<char*>(&argv[i][string_start]);
int length = static_cast<int>(strlen(string_ref));
if (!STRNCASECMP(string_argv, string_ref, length)) {
*string_retval = &string_argv[length + 1];
bFound = true;
continue;
}
}
}
if (!bFound) {
*string_retval = NULL;
}
return bFound;
}
//////////////////////////////////////////////////////////////////////////////
//! Find the path for a file assuming that
//! files are found in the searchPath.
//!
//! @return the path if succeeded, otherwise 0
//! @param filename name of the file
//! @param executable_path optional absolute path of the executable
//////////////////////////////////////////////////////////////////////////////
inline char *sdkFindFilePath(const char *filename,
const char *executable_path) {
// <executable_name> defines a variable that is replaced with the name of the
// executable
// Typical relative search paths to locate needed companion files (e.g. sample
// input data, or JIT source files) The origin for the relative search may be
// the .exe file, a .bat file launching an .exe, a browser .exe launching the
// .exe or .bat, etc
const char *searchPath[] = {
"./", // same dir
"./<executable_name>_data_files/",
"./common/", // "/common/" subdir
"./common/data/", // "/common/data/" subdir
"./data/", // "/data/" subdir
"./src/", // "/src/" subdir
"./src/<executable_name>/data/", // "/src/<executable_name>/data/" subdir
"./inc/", // "/inc/" subdir
"./0_Simple/", // "/0_Simple/" subdir
"./1_Utilities/", // "/1_Utilities/" subdir
"./2_Graphics/", // "/2_Graphics/" subdir
"./3_Imaging/", // "/3_Imaging/" subdir
"./4_Finance/", // "/4_Finance/" subdir
"./5_Simulations/", // "/5_Simulations/" subdir
"./6_Advanced/", // "/6_Advanced/" subdir
"./7_CUDALibraries/", // "/7_CUDALibraries/" subdir
"./8_Android/", // "/8_Android/" subdir
"./samples/", // "/samples/" subdir
"./0_Simple/<executable_name>/data/", // "/0_Simple/<executable_name>/data/"
// subdir
"./1_Utilities/<executable_name>/data/", // "/1_Utilities/<executable_name>/data/"
// subdir
"./2_Graphics/<executable_name>/data/", // "/2_Graphics/<executable_name>/data/"
// subdir
"./3_Imaging/<executable_name>/data/", // "/3_Imaging/<executable_name>/data/"
// subdir
"./4_Finance/<executable_name>/data/", // "/4_Finance/<executable_name>/data/"
// subdir
"./5_Simulations/<executable_name>/data/", // "/5_Simulations/<executable_name>/data/"
// subdir
"./6_Advanced/<executable_name>/data/", // "/6_Advanced/<executable_name>/data/"
// subdir
"./7_CUDALibraries/<executable_name>/", // "/7_CUDALibraries/<executable_name>/"
// subdir
"./7_CUDALibraries/<executable_name>/data/", // "/7_CUDALibraries/<executable_name>/data/"
// subdir
"../", // up 1 in tree
"../common/", // up 1 in tree, "/common/" subdir
"../common/data/", // up 1 in tree, "/common/data/" subdir
"../data/", // up 1 in tree, "/data/" subdir
"../src/", // up 1 in tree, "/src/" subdir
"../inc/", // up 1 in tree, "/inc/" subdir
"../0_Simple/<executable_name>/data/", // up 1 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../1_Utilities/<executable_name>/data/", // up 1 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../2_Graphics/<executable_name>/data/", // up 1 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../3_Imaging/<executable_name>/data/", // up 1 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../4_Finance/<executable_name>/data/", // up 1 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../5_Simulations/<executable_name>/data/", // up 1 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../6_Advanced/<executable_name>/data/", // up 1 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../7_CUDALibraries/<executable_name>/data/", // up 1 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../8_Android/<executable_name>/data/", // up 1 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../samples/<executable_name>/data/", // up 1 in tree,
// "/samples/<executable_name>/"
// subdir
"../../", // up 2 in tree
"../../common/", // up 2 in tree, "/common/" subdir
"../../common/data/", // up 2 in tree, "/common/data/" subdir
"../../data/", // up 2 in tree, "/data/" subdir
"../../src/", // up 2 in tree, "/src/" subdir
"../../inc/", // up 2 in tree, "/inc/" subdir
"../../sandbox/<executable_name>/data/", // up 2 in tree,
// "/sandbox/<executable_name>/"
// subdir
"../../0_Simple/<executable_name>/data/", // up 2 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../1_Utilities/<executable_name>/data/", // up 2 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../2_Graphics/<executable_name>/data/", // up 2 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../3_Imaging/<executable_name>/data/", // up 2 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../4_Finance/<executable_name>/data/", // up 2 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../5_Simulations/<executable_name>/data/", // up 2 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../6_Advanced/<executable_name>/data/", // up 2 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../7_CUDALibraries/<executable_name>/data/", // up 2 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../8_Android/<executable_name>/data/", // up 2 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../samples/<executable_name>/data/", // up 2 in tree,
// "/samples/<executable_name>/"
// subdir
"../../../", // up 3 in tree
"../../../src/<executable_name>/", // up 3 in tree,
// "/src/<executable_name>/" subdir
"../../../src/<executable_name>/data/", // up 3 in tree,
// "/src/<executable_name>/data/"
// subdir
"../../../src/<executable_name>/src/", // up 3 in tree,
// "/src/<executable_name>/src/"
// subdir
"../../../src/<executable_name>/inc/", // up 3 in tree,
// "/src/<executable_name>/inc/"
// subdir
"../../../sandbox/<executable_name>/", // up 3 in tree,
// "/sandbox/<executable_name>/"
// subdir
"../../../sandbox/<executable_name>/data/", // up 3 in tree,
// "/sandbox/<executable_name>/data/"
// subdir
"../../../sandbox/<executable_name>/src/", // up 3 in tree,
// "/sandbox/<executable_name>/src/"
// subdir
"../../../sandbox/<executable_name>/inc/", // up 3 in tree,
// "/sandbox/<executable_name>/inc/"
// subdir
"../../../0_Simple/<executable_name>/data/", // up 3 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../../1_Utilities/<executable_name>/data/", // up 3 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../../2_Graphics/<executable_name>/data/", // up 3 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../../3_Imaging/<executable_name>/data/", // up 3 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../../4_Finance/<executable_name>/data/", // up 3 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../../5_Simulations/<executable_name>/data/", // up 3 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../../6_Advanced/<executable_name>/data/", // up 3 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../../7_CUDALibraries/<executable_name>/data/", // up 3 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../../8_Android/<executable_name>/data/", // up 3 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../../0_Simple/<executable_name>/", // up 3 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../../1_Utilities/<executable_name>/", // up 3 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../../2_Graphics/<executable_name>/", // up 3 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../../3_Imaging/<executable_name>/", // up 3 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../../4_Finance/<executable_name>/", // up 3 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../../5_Simulations/<executable_name>/", // up 3 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../../6_Advanced/<executable_name>/", // up 3 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../../7_CUDALibraries/<executable_name>/", // up 3 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../../8_Android/<executable_name>/", // up 3 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../../samples/<executable_name>/data/", // up 3 in tree,
// "/samples/<executable_name>/"
// subdir
"../../../common/", // up 3 in tree, "../../../common/" subdir
"../../../common/data/", // up 3 in tree, "../../../common/data/" subdir
"../../../data/", // up 3 in tree, "../../../data/" subdir
"../../../../", // up 4 in tree
"../../../../src/<executable_name>/", // up 4 in tree,
// "/src/<executable_name>/" subdir
"../../../../src/<executable_name>/data/", // up 4 in tree,
// "/src/<executable_name>/data/"
// subdir
"../../../../src/<executable_name>/src/", // up 4 in tree,
// "/src/<executable_name>/src/"
// subdir
"../../../../src/<executable_name>/inc/", // up 4 in tree,
// "/src/<executable_name>/inc/"
// subdir
"../../../../sandbox/<executable_name>/", // up 4 in tree,
// "/sandbox/<executable_name>/"
// subdir
"../../../../sandbox/<executable_name>/data/", // up 4 in tree,
// "/sandbox/<executable_name>/data/"
// subdir
"../../../../sandbox/<executable_name>/src/", // up 4 in tree,
// "/sandbox/<executable_name>/src/"
// subdir
"../../../../sandbox/<executable_name>/inc/", // up 4 in tree,
// "/sandbox/<executable_name>/inc/"
// subdir
"../../../../0_Simple/<executable_name>/data/", // up 4 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../../../1_Utilities/<executable_name>/data/", // up 4 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../../../2_Graphics/<executable_name>/data/", // up 4 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../../../3_Imaging/<executable_name>/data/", // up 4 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../../../4_Finance/<executable_name>/data/", // up 4 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../../../5_Simulations/<executable_name>/data/", // up 4 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../../../6_Advanced/<executable_name>/data/", // up 4 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../../../7_CUDALibraries/<executable_name>/data/", // up 4 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../../../8_Android/<executable_name>/data/", // up 4 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../../../0_Simple/<executable_name>/", // up 4 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../../../1_Utilities/<executable_name>/", // up 4 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../../../2_Graphics/<executable_name>/", // up 4 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../../../3_Imaging/<executable_name>/", // up 4 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../../../4_Finance/<executable_name>/", // up 4 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../../../5_Simulations/<executable_name>/", // up 4 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../../../6_Advanced/<executable_name>/", // up 4 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../../../7_CUDALibraries/<executable_name>/", // up 4 in tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../../../8_Android/<executable_name>/", // up 4 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../../../samples/<executable_name>/data/", // up 4 in tree,
// "/samples/<executable_name>/"
// subdir
"../../../../common/", // up 4 in tree, "../../../common/" subdir
"../../../../common/data/", // up 4 in tree, "../../../common/data/"
// subdir
"../../../../data/", // up 4 in tree, "../../../data/" subdir
"../../../../../", // up 5 in tree
"../../../../../src/<executable_name>/", // up 5 in tree,
// "/src/<executable_name>/"
// subdir
"../../../../../src/<executable_name>/data/", // up 5 in tree,
// "/src/<executable_name>/data/"
// subdir
"../../../../../src/<executable_name>/src/", // up 5 in tree,
// "/src/<executable_name>/src/"
// subdir
"../../../../../src/<executable_name>/inc/", // up 5 in tree,
// "/src/<executable_name>/inc/"
// subdir
"../../../../../sandbox/<executable_name>/", // up 5 in tree,
// "/sandbox/<executable_name>/"
// subdir
"../../../../../sandbox/<executable_name>/data/", // up 5 in tree,
// "/sandbox/<executable_name>/data/"
// subdir
"../../../../../sandbox/<executable_name>/src/", // up 5 in tree,
// "/sandbox/<executable_name>/src/"
// subdir
"../../../../../sandbox/<executable_name>/inc/", // up 5 in tree,
// "/sandbox/<executable_name>/inc/"
// subdir
"../../../../../0_Simple/<executable_name>/data/", // up 5 in tree,
// "/0_Simple/<executable_name>/"
// subdir
"../../../../../1_Utilities/<executable_name>/data/", // up 5 in tree,
// "/1_Utilities/<executable_name>/"
// subdir
"../../../../../2_Graphics/<executable_name>/data/", // up 5 in tree,
// "/2_Graphics/<executable_name>/"
// subdir
"../../../../../3_Imaging/<executable_name>/data/", // up 5 in tree,
// "/3_Imaging/<executable_name>/"
// subdir
"../../../../../4_Finance/<executable_name>/data/", // up 5 in tree,
// "/4_Finance/<executable_name>/"
// subdir
"../../../../../5_Simulations/<executable_name>/data/", // up 5 in tree,
// "/5_Simulations/<executable_name>/"
// subdir
"../../../../../6_Advanced/<executable_name>/data/", // up 5 in tree,
// "/6_Advanced/<executable_name>/"
// subdir
"../../../../../7_CUDALibraries/<executable_name>/data/", // up 5 in
// tree,
// "/7_CUDALibraries/<executable_name>/"
// subdir
"../../../../../8_Android/<executable_name>/data/", // up 5 in tree,
// "/8_Android/<executable_name>/"
// subdir
"../../../../../samples/<executable_name>/data/", // up 5 in tree,
// "/samples/<executable_name>/"
// subdir
"../../../../../common/", // up 5 in tree, "../../../common/" subdir
"../../../../../common/data/", // up 5 in tree, "../../../common/data/"
// subdir
};
// Extract the executable name
std::string executable_name;
if (executable_path != 0) {
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
}
// Loop over all search paths and return the first hit
for (unsigned int i = 0; i < sizeof(searchPath) / sizeof(char *); ++i) {
std::string path(searchPath[i]);
size_t executable_name_pos = path.find("<executable_name>");
// If there is executable_name variable in the searchPath
// replace it with the value
if (executable_name_pos != std::string::npos) {
if (executable_path != 0) {
path.replace(executable_name_pos, strlen("<executable_name>"),
executable_name);
} else {
// Skip this path entry if no executable argument is given
continue;
}
}
#ifdef _DEBUG
printf("sdkFindFilePath <%s> in %s\n", filename, path.c_str());
#endif
// Test if the file exists
path.append(filename);
FILE *fp;
FOPEN(fp, path.c_str(), "rb");
if (fp != NULL) {
fclose(fp);
// File found
// returning an allocated array here for backwards compatibility reasons
char *file_path = reinterpret_cast<char *>(malloc(path.length() + 1));
STRCPY(file_path, path.length() + 1, path.c_str());
return file_path;
}
if (fp) {
fclose(fp);
}
}
// File not found
return 0;
}
#endif // COMMON_HELPER_STRING_H_

465
Common/helper_timer.h Normal file
View File

@@ -0,0 +1,465 @@
/* Copyright (c) 2018, 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.
*/
// Helper Timing Functions
#ifndef COMMON_HELPER_TIMER_H_
#define COMMON_HELPER_TIMER_H_
#ifndef EXIT_WAIVED
#define EXIT_WAIVED 2
#endif
// includes, system
#include <vector>
// includes, project
#include <exception.h>
// Definition of the StopWatch Interface, this is used if we don't want to use
// the CUT functions But rather in a self contained class interface
class StopWatchInterface {
public:
StopWatchInterface() {}
virtual ~StopWatchInterface() {}
public:
//! Start time measurement
virtual void start() = 0;
//! Stop time measurement
virtual void stop() = 0;
//! Reset time counters to zero
virtual void reset() = 0;
//! Time in msec. after start. If the stop watch is still running (i.e. there
//! was no call to stop()) then the elapsed time is returned, otherwise the
//! time between the last start() and stop call is returned
virtual float getTime() = 0;
//! Mean time to date based on the number of times the stopwatch has been
//! _stopped_ (ie finished sessions) and the current total time
virtual float getAverageTime() = 0;
};
//////////////////////////////////////////////////////////////////
// Begin Stopwatch timer class definitions for all OS platforms //
//////////////////////////////////////////////////////////////////
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
// includes, system
#define WINDOWS_LEAN_AND_MEAN
#include <windows.h>
#undef min
#undef max
//! Windows specific implementation of StopWatch
class StopWatchWin : public StopWatchInterface {
public:
//! Constructor, default
StopWatchWin()
: start_time(),
end_time(),
diff_time(0.0f),
total_time(0.0f),
running(false),
clock_sessions(0),
freq(0),
freq_set(false) {
if (!freq_set) {
// helper variable
LARGE_INTEGER temp;
// get the tick frequency from the OS
QueryPerformanceFrequency(reinterpret_cast<LARGE_INTEGER *>(&temp));
// convert to type in which it is needed
freq = (static_cast<double>(temp.QuadPart)) / 1000.0;
// rememeber query
freq_set = true;
}
}
// Destructor
~StopWatchWin() {}
public:
//! Start time measurement
inline void start();
//! Stop time measurement
inline void stop();
//! Reset time counters to zero
inline void reset();
//! Time in msec. after start. If the stop watch is still running (i.e. there
//! was no call to stop()) then the elapsed time is returned, otherwise the
//! time between the last start() and stop call is returned
inline float getTime();
//! Mean time to date based on the number of times the stopwatch has been
//! _stopped_ (ie finished sessions) and the current total time
inline float getAverageTime();
private:
// member variables
//! Start of measurement
LARGE_INTEGER start_time;
//! End of measurement
LARGE_INTEGER end_time;
//! Time difference between the last start and stop
float diff_time;
//! TOTAL time difference between starts and stops
float total_time;
//! flag if the stop watch is running
bool running;
//! Number of times clock has been started
//! and stopped to allow averaging
int clock_sessions;
//! tick frequency
double freq;
//! flag if the frequency has been set
bool freq_set;
};
// functions, inlined
////////////////////////////////////////////////////////////////////////////////
//! Start time measurement
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchWin::start() {
QueryPerformanceCounter(reinterpret_cast<LARGE_INTEGER *>(&start_time));
running = true;
}
////////////////////////////////////////////////////////////////////////////////
//! Stop time measurement and increment add to the current diff_time summation
//! variable. Also increment the number of times this clock has been run.
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchWin::stop() {
QueryPerformanceCounter(reinterpret_cast<LARGE_INTEGER *>(&end_time));
diff_time = static_cast<float>(((static_cast<double>(end_time.QuadPart) -
static_cast<double>(start_time.QuadPart)) /
freq));
total_time += diff_time;
clock_sessions++;
running = false;
}
////////////////////////////////////////////////////////////////////////////////
//! Reset the timer to 0. Does not change the timer running state but does
//! recapture this point in time as the current start time if it is running.
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchWin::reset() {
diff_time = 0;
total_time = 0;
clock_sessions = 0;
if (running) {
QueryPerformanceCounter(reinterpret_cast<LARGE_INTEGER *>(&start_time));
}
}
////////////////////////////////////////////////////////////////////////////////
//! Time in msec. after start. If the stop watch is still running (i.e. there
//! was no call to stop()) then the elapsed time is returned added to the
//! current diff_time sum, otherwise the current summed time difference alone
//! is returned.
////////////////////////////////////////////////////////////////////////////////
inline float StopWatchWin::getTime() {
// Return the TOTAL time to date
float retval = total_time;
if (running) {
LARGE_INTEGER temp;
QueryPerformanceCounter(reinterpret_cast<LARGE_INTEGER *>(&temp));
retval += static_cast<float>(((static_cast<double>(temp.QuadPart) -
static_cast<double>(start_time.QuadPart)) /
freq));
}
return retval;
}
////////////////////////////////////////////////////////////////////////////////
//! Time in msec. for a single run based on the total number of COMPLETED runs
//! and the total time.
////////////////////////////////////////////////////////////////////////////////
inline float StopWatchWin::getAverageTime() {
return (clock_sessions > 0) ? (total_time / clock_sessions) : 0.0f;
}
#else
// Declarations for Stopwatch on Linux and Mac OSX
// includes, system
#include <sys/time.h>
#include <ctime>
//! Windows specific implementation of StopWatch
class StopWatchLinux : public StopWatchInterface {
public:
//! Constructor, default
StopWatchLinux()
: start_time(),
diff_time(0.0),
total_time(0.0),
running(false),
clock_sessions(0) {}
// Destructor
virtual ~StopWatchLinux() {}
public:
//! Start time measurement
inline void start();
//! Stop time measurement
inline void stop();
//! Reset time counters to zero
inline void reset();
//! Time in msec. after start. If the stop watch is still running (i.e. there
//! was no call to stop()) then the elapsed time is returned, otherwise the
//! time between the last start() and stop call is returned
inline float getTime();
//! Mean time to date based on the number of times the stopwatch has been
//! _stopped_ (ie finished sessions) and the current total time
inline float getAverageTime();
private:
// helper functions
//! Get difference between start time and current time
inline float getDiffTime();
private:
// member variables
//! Start of measurement
struct timeval start_time;
//! Time difference between the last start and stop
float diff_time;
//! TOTAL time difference between starts and stops
float total_time;
//! flag if the stop watch is running
bool running;
//! Number of times clock has been started
//! and stopped to allow averaging
int clock_sessions;
};
// functions, inlined
////////////////////////////////////////////////////////////////////////////////
//! Start time measurement
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchLinux::start() {
gettimeofday(&start_time, 0);
running = true;
}
////////////////////////////////////////////////////////////////////////////////
//! Stop time measurement and increment add to the current diff_time summation
//! variable. Also increment the number of times this clock has been run.
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchLinux::stop() {
diff_time = getDiffTime();
total_time += diff_time;
running = false;
clock_sessions++;
}
////////////////////////////////////////////////////////////////////////////////
//! Reset the timer to 0. Does not change the timer running state but does
//! recapture this point in time as the current start time if it is running.
////////////////////////////////////////////////////////////////////////////////
inline void StopWatchLinux::reset() {
diff_time = 0;
total_time = 0;
clock_sessions = 0;
if (running) {
gettimeofday(&start_time, 0);
}
}
////////////////////////////////////////////////////////////////////////////////
//! Time in msec. after start. If the stop watch is still running (i.e. there
//! was no call to stop()) then the elapsed time is returned added to the
//! current diff_time sum, otherwise the current summed time difference alone
//! is returned.
////////////////////////////////////////////////////////////////////////////////
inline float StopWatchLinux::getTime() {
// Return the TOTAL time to date
float retval = total_time;
if (running) {
retval += getDiffTime();
}
return retval;
}
////////////////////////////////////////////////////////////////////////////////
//! Time in msec. for a single run based on the total number of COMPLETED runs
//! and the total time.
////////////////////////////////////////////////////////////////////////////////
inline float StopWatchLinux::getAverageTime() {
return (clock_sessions > 0) ? (total_time / clock_sessions) : 0.0f;
}
////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////
inline float StopWatchLinux::getDiffTime() {
struct timeval t_time;
gettimeofday(&t_time, 0);
// time difference in milli-seconds
return static_cast<float>(1000.0 * (t_time.tv_sec - start_time.tv_sec) +
(0.001 * (t_time.tv_usec - start_time.tv_usec)));
}
#endif // WIN32
////////////////////////////////////////////////////////////////////////////////
//! Timer functionality exported
////////////////////////////////////////////////////////////////////////////////
//! Create a new timer
//! @return true if a time has been created, otherwise false
//! @param name of the new timer, 0 if the creation failed
////////////////////////////////////////////////////////////////////////////////
inline bool sdkCreateTimer(StopWatchInterface **timer_interface) {
// printf("sdkCreateTimer called object %08x\n", (void *)*timer_interface);
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
*timer_interface = reinterpret_cast<StopWatchInterface *>(new StopWatchWin());
#else
*timer_interface =
reinterpret_cast<StopWatchInterface *>(new StopWatchLinux());
#endif
return (*timer_interface != NULL) ? true : false;
}
////////////////////////////////////////////////////////////////////////////////
//! Delete a timer
//! @return true if a time has been deleted, otherwise false
//! @param name of the timer to delete
////////////////////////////////////////////////////////////////////////////////
inline bool sdkDeleteTimer(StopWatchInterface **timer_interface) {
// printf("sdkDeleteTimer called object %08x\n", (void *)*timer_interface);
if (*timer_interface) {
delete *timer_interface;
*timer_interface = NULL;
}
return true;
}
////////////////////////////////////////////////////////////////////////////////
//! Start the time with name \a name
//! @param name name of the timer to start
////////////////////////////////////////////////////////////////////////////////
inline bool sdkStartTimer(StopWatchInterface **timer_interface) {
// printf("sdkStartTimer called object %08x\n", (void *)*timer_interface);
if (*timer_interface) {
(*timer_interface)->start();
}
return true;
}
////////////////////////////////////////////////////////////////////////////////
//! Stop the time with name \a name. Does not reset.
//! @param name name of the timer to stop
////////////////////////////////////////////////////////////////////////////////
inline bool sdkStopTimer(StopWatchInterface **timer_interface) {
// printf("sdkStopTimer called object %08x\n", (void *)*timer_interface);
if (*timer_interface) {
(*timer_interface)->stop();
}
return true;
}
////////////////////////////////////////////////////////////////////////////////
//! Resets the timer's counter.
//! @param name name of the timer to reset.
////////////////////////////////////////////////////////////////////////////////
inline bool sdkResetTimer(StopWatchInterface **timer_interface) {
// printf("sdkResetTimer called object %08x\n", (void *)*timer_interface);
if (*timer_interface) {
(*timer_interface)->reset();
}
return true;
}
////////////////////////////////////////////////////////////////////////////////
//! Return the average time for timer execution as the total time
//! for the timer dividied by the number of completed (stopped) runs the timer
//! has made.
//! Excludes the current running time if the timer is currently running.
//! @param name name of the timer to return the time of
////////////////////////////////////////////////////////////////////////////////
inline float sdkGetAverageTimerValue(StopWatchInterface **timer_interface) {
// printf("sdkGetAverageTimerValue called object %08x\n", (void
// *)*timer_interface);
if (*timer_interface) {
return (*timer_interface)->getAverageTime();
} else {
return 0.0f;
}
}
////////////////////////////////////////////////////////////////////////////////
//! Total execution time for the timer over all runs since the last reset
//! or timer creation.
//! @param name name of the timer to obtain the value of.
////////////////////////////////////////////////////////////////////////////////
inline float sdkGetTimerValue(StopWatchInterface **timer_interface) {
// printf("sdkGetTimerValue called object %08x\n", (void *)*timer_interface);
if (*timer_interface) {
return (*timer_interface)->getTime();
} else {
return 0.0f;
}
}
#endif // COMMON_HELPER_TIMER_H_

170
Common/nvrtc_helper.h Normal file
View File

@@ -0,0 +1,170 @@
/* Copyright (c) 2018, 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 COMMON_NVRTC_HELPER_H_
#define COMMON_NVRTC_HELPER_H_ 1
#include <cuda.h>
#include <helper_cuda_drvapi.h>
#include <nvrtc.h>
#include <fstream>
#include <iostream>
#include <sstream>
#include <string>
#define NVRTC_SAFE_CALL(Name, x) \
do { \
nvrtcResult result = x; \
if (result != NVRTC_SUCCESS) { \
std::cerr << "\nerror: " << Name << " failed with error " \
<< nvrtcGetErrorString(result); \
exit(1); \
} \
} while (0)
void compileFileToPTX(char *filename, int argc, char **argv, char **ptxResult,
size_t *ptxResultSize, int requiresCGheaders) {
std::ifstream inputFile(filename,
std::ios::in | std::ios::binary | std::ios::ate);
if (!inputFile.is_open()) {
std::cerr << "\nerror: unable to open " << filename << " for reading!\n";
exit(1);
}
std::streampos pos = inputFile.tellg();
size_t inputSize = (size_t)pos;
char *memBlock = new char[inputSize + 1];
inputFile.seekg(0, std::ios::beg);
inputFile.read(memBlock, inputSize);
inputFile.close();
memBlock[inputSize] = '\x0';
int numCompileOptions = 0;
char *compileParams[1];
if (requiresCGheaders) {
std::string compileOptions;
char HeaderNames[256];
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
sprintf_s(HeaderNames, sizeof(HeaderNames), "%s", "cooperative_groups.h");
#else
snprintf(HeaderNames, sizeof(HeaderNames), "%s", "cooperative_groups.h");
#endif
compileOptions = "--include-path=";
std::string path = sdkFindFilePath(HeaderNames, argv[0]);
if (!path.empty()) {
std::size_t found = path.find(HeaderNames);
path.erase(found);
} else {
printf(
"\nCooperativeGroups headers not found, please install it in %s "
"sample directory..\n Exiting..\n",
argv[0]);
}
compileOptions += path.c_str();
compileParams[0] = reinterpret_cast<char *>(
malloc(sizeof(char) * (compileOptions.length() + 1)));
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
sprintf_s(compileParams[0], sizeof(char) * (compileOptions.length() + 1),
"%s", compileOptions.c_str());
#else
snprintf(compileParams[0], compileOptions.size(), "%s",
compileOptions.c_str());
#endif
numCompileOptions++;
}
// compile
nvrtcProgram prog;
NVRTC_SAFE_CALL("nvrtcCreateProgram",
nvrtcCreateProgram(&prog, memBlock, filename, 0, NULL, NULL));
nvrtcResult res = nvrtcCompileProgram(prog, numCompileOptions, compileParams);
// dump log
size_t logSize;
NVRTC_SAFE_CALL("nvrtcGetProgramLogSize",
nvrtcGetProgramLogSize(prog, &logSize));
char *log = reinterpret_cast<char *>(malloc(sizeof(char) * logSize + 1));
NVRTC_SAFE_CALL("nvrtcGetProgramLog", nvrtcGetProgramLog(prog, log));
log[logSize] = '\x0';
if (strlen(log) >= 2) {
std::cerr << "\n compilation log ---\n";
std::cerr << log;
std::cerr << "\n end log ---\n";
}
free(log);
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]);
}
CUmodule loadPTX(char *ptx, int argc, char **argv) {
CUmodule module;
CUcontext context;
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));
checkCudaErrors(cuDeviceGetName(deviceName, 256, cuDevice));
printf("> GPU Device has SM %d.%d compute capability\n", major, minor);
checkCudaErrors(cuInit(0));
checkCudaErrors(cuDeviceGet(&cuDevice, 0));
checkCudaErrors(cuCtxCreate(&context, 0, cuDevice));
checkCudaErrors(cuModuleLoadDataEx(&module, ptx, 0, 0, 0));
free(ptx);
return module;
}
#endif // COMMON_NVRTC_HELPER_H_