* Copyright 1993-2017 NVIDIA Corporation. All rights reserved.
* Please refer to the NVIDIA end user license agreement (EULA) associated
* with this source code for terms and conditions that govern your use of
* this software. Any use, reproduction, disclosure, or distribution of
Copyright By PowCoder代写 加微信 powcoder
* this software and related documentation outside the terms of the EULA
* is strictly prohibited.
////////////////////////////////////////////////////////////////////////////////
// These are CUDA Helper functions for initialization and error checking
#ifndef COMMON_HELPER_CUDA_H_
#define COMMON_HELPER_CUDA_H_
#pragma once
#include
#include
#include
#include
#include
#ifndef EXIT_WAIVED
#define EXIT_WAIVED 2
// Note, it is required that your SDK sample to include the proper header
// files, please refer the CUDA examples for examples of the needed CUDA
// headers, which may change depending on which CUDA functions are used.
// CUDA Runtime error messages
#ifdef __DRIVER_TYPES_H__
static const char *_cudaGetErrorEnum(cudaError_t error) {
return cudaGetErrorName(error);
#ifdef CUDA_DRIVER_API
// CUDA Driver API errors
static const char *_cudaGetErrorEnum(CUresult error) {
static char unknown[] = “
const char *ret = NULL;
cuGetErrorName(error, &ret);
return ret ? ret : unknown;
#ifdef CUBLAS_API_H_
// cuBLAS API errors
static const char *_cudaGetErrorEnum(cublasStatus_t error) {
switch (error) {
case CUBLAS_STATUS_SUCCESS:
return “CUBLAS_STATUS_SUCCESS”;
case CUBLAS_STATUS_NOT_INITIALIZED:
return “CUBLAS_STATUS_NOT_INITIALIZED”;
case CUBLAS_STATUS_ALLOC_FAILED:
return “CUBLAS_STATUS_ALLOC_FAILED”;
case CUBLAS_STATUS_INVALID_VALUE:
return “CUBLAS_STATUS_INVALID_VALUE”;
case CUBLAS_STATUS_ARCH_MISMATCH:
return “CUBLAS_STATUS_ARCH_MISMATCH”;
case CUBLAS_STATUS_MAPPING_ERROR:
return “CUBLAS_STATUS_MAPPING_ERROR”;
case CUBLAS_STATUS_EXECUTION_FAILED:
return “CUBLAS_STATUS_EXECUTION_FAILED”;
case CUBLAS_STATUS_INTERNAL_ERROR:
return “CUBLAS_STATUS_INTERNAL_ERROR”;
case CUBLAS_STATUS_NOT_SUPPORTED:
return “CUBLAS_STATUS_NOT_SUPPORTED”;
case CUBLAS_STATUS_LICENSE_ERROR:
return “CUBLAS_STATUS_LICENSE_ERROR”;
return “
#ifdef _CUFFT_H_
// cuFFT API errors
static const char *_cudaGetErrorEnum(cufftResult error) {
switch (error) {
case CUFFT_SUCCESS:
return “CUFFT_SUCCESS”;
case CUFFT_INVALID_PLAN:
return “CUFFT_INVALID_PLAN”;
case CUFFT_ALLOC_FAILED:
return “CUFFT_ALLOC_FAILED”;
case CUFFT_INVALID_TYPE:
return “CUFFT_INVALID_TYPE”;
case CUFFT_INVALID_VALUE:
return “CUFFT_INVALID_VALUE”;
case CUFFT_INTERNAL_ERROR:
return “CUFFT_INTERNAL_ERROR”;
case CUFFT_EXEC_FAILED:
return “CUFFT_EXEC_FAILED”;
case CUFFT_SETUP_FAILED:
return “CUFFT_SETUP_FAILED”;
case CUFFT_INVALID_SIZE:
return “CUFFT_INVALID_SIZE”;
case CUFFT_UNALIGNED_DATA:
return “CUFFT_UNALIGNED_DATA”;
case CUFFT_INCOMPLETE_PARAMETER_LIST:
return “CUFFT_INCOMPLETE_PARAMETER_LIST”;
case CUFFT_INVALID_DEVICE:
return “CUFFT_INVALID_DEVICE”;
case CUFFT_PARSE_ERROR:
return “CUFFT_PARSE_ERROR”;
case CUFFT_NO_WORKSPACE:
return “CUFFT_NO_WORKSPACE”;
case CUFFT_NOT_IMPLEMENTED:
return “CUFFT_NOT_IMPLEMENTED”;
case CUFFT_LICENSE_ERROR:
return “CUFFT_LICENSE_ERROR”;
case CUFFT_NOT_SUPPORTED:
return “CUFFT_NOT_SUPPORTED”;
return “
#ifdef CUSPARSEAPI
// cuSPARSE API errors
static const char *_cudaGetErrorEnum(cusparseStatus_t error) {
switch (error) {
case CUSPARSE_STATUS_SUCCESS:
return “CUSPARSE_STATUS_SUCCESS”;
case CUSPARSE_STATUS_NOT_INITIALIZED:
return “CUSPARSE_STATUS_NOT_INITIALIZED”;
case CUSPARSE_STATUS_ALLOC_FAILED:
return “CUSPARSE_STATUS_ALLOC_FAILED”;
case CUSPARSE_STATUS_INVALID_VALUE:
return “CUSPARSE_STATUS_INVALID_VALUE”;
case CUSPARSE_STATUS_ARCH_MISMATCH:
return “CUSPARSE_STATUS_ARCH_MISMATCH”;
case CUSPARSE_STATUS_MAPPING_ERROR:
return “CUSPARSE_STATUS_MAPPING_ERROR”;
case CUSPARSE_STATUS_EXECUTION_FAILED:
return “CUSPARSE_STATUS_EXECUTION_FAILED”;
case CUSPARSE_STATUS_INTERNAL_ERROR:
return “CUSPARSE_STATUS_INTERNAL_ERROR”;
case CUSPARSE_STATUS_MATRIX_TYPE_NOT_SUPPORTED:
return “CUSPARSE_STATUS_MATRIX_TYPE_NOT_SUPPORTED”;
return “
#ifdef CUSOLVER_COMMON_H_
// cuSOLVER API errors
static const char *_cudaGetErrorEnum(cusolverStatus_t error) {
switch (error) {
case CUSOLVER_STATUS_SUCCESS:
return “CUSOLVER_STATUS_SUCCESS”;
case CUSOLVER_STATUS_NOT_INITIALIZED:
return “CUSOLVER_STATUS_NOT_INITIALIZED”;
case CUSOLVER_STATUS_ALLOC_FAILED:
return “CUSOLVER_STATUS_ALLOC_FAILED”;
case CUSOLVER_STATUS_INVALID_VALUE:
return “CUSOLVER_STATUS_INVALID_VALUE”;
case CUSOLVER_STATUS_ARCH_MISMATCH:
return “CUSOLVER_STATUS_ARCH_MISMATCH”;
case CUSOLVER_STATUS_MAPPING_ERROR:
return “CUSOLVER_STATUS_MAPPING_ERROR”;
case CUSOLVER_STATUS_EXECUTION_FAILED:
return “CUSOLVER_STATUS_EXECUTION_FAILED”;
case CUSOLVER_STATUS_INTERNAL_ERROR:
return “CUSOLVER_STATUS_INTERNAL_ERROR”;
case CUSOLVER_STATUS_MATRIX_TYPE_NOT_SUPPORTED:
return “CUSOLVER_STATUS_MATRIX_TYPE_NOT_SUPPORTED”;
case CUSOLVER_STATUS_NOT_SUPPORTED:
return “CUSOLVER_STATUS_NOT_SUPPORTED “;
case CUSOLVER_STATUS_ZERO_PIVOT:
return “CUSOLVER_STATUS_ZERO_PIVOT”;
case CUSOLVER_STATUS_INVALID_LICENSE:
return “CUSOLVER_STATUS_INVALID_LICENSE”;
return “
#ifdef CURAND_H_
// cuRAND API errors
static const char *_cudaGetErrorEnum(curandStatus_t error) {
switch (error) {
case CURAND_STATUS_SUCCESS:
return “CURAND_STATUS_SUCCESS”;
case CURAND_STATUS_VERSION_MISMATCH:
return “CURAND_STATUS_VERSION_MISMATCH”;
case CURAND_STATUS_NOT_INITIALIZED:
return “CURAND_STATUS_NOT_INITIALIZED”;
case CURAND_STATUS_ALLOCATION_FAILED:
return “CURAND_STATUS_ALLOCATION_FAILED”;
case CURAND_STATUS_TYPE_ERROR:
return “CURAND_STATUS_TYPE_ERROR”;
case CURAND_STATUS_OUT_OF_RANGE:
return “CURAND_STATUS_OUT_OF_RANGE”;
case CURAND_STATUS_LENGTH_NOT_MULTIPLE:
return “CURAND_STATUS_LENGTH_NOT_MULTIPLE”;
case CURAND_STATUS_DOUBLE_PRECISION_REQUIRED:
return “CURAND_STATUS_DOUBLE_PRECISION_REQUIRED”;
case CURAND_STATUS_LAUNCH_FAILURE:
return “CURAND_STATUS_LAUNCH_FAILURE”;
case CURAND_STATUS_PREEXISTING_FAILURE:
return “CURAND_STATUS_PREEXISTING_FAILURE”;
case CURAND_STATUS_INITIALIZATION_FAILED:
return “CURAND_STATUS_INITIALIZATION_FAILED”;
case CURAND_STATUS_ARCH_MISMATCH:
return “CURAND_STATUS_ARCH_MISMATCH”;
case CURAND_STATUS_INTERNAL_ERROR:
return “CURAND_STATUS_INTERNAL_ERROR”;
return “
#ifdef NVJPEGAPI
// nvJPEG API errors
static const char *_cudaGetErrorEnum(nvjpegStatus_t error) {
switch (error) {
case NVJPEG_STATUS_SUCCESS:
return “NVJPEG_STATUS_SUCCESS”;
case NVJPEG_STATUS_NOT_INITIALIZED:
return “NVJPEG_STATUS_NOT_INITIALIZED”;
case NVJPEG_STATUS_INVALID_PARAMETER:
return “NVJPEG_STATUS_INVALID_PARAMETER”;
case NVJPEG_STATUS_BAD_JPEG:
return “NVJPEG_STATUS_BAD_JPEG”;
case NVJPEG_STATUS_JPEG_NOT_SUPPORTED:
return “NVJPEG_STATUS_JPEG_NOT_SUPPORTED”;
case NVJPEG_STATUS_ALLOCATOR_FAILURE:
return “NVJPEG_STATUS_ALLOCATOR_FAILURE”;
case NVJPEG_STATUS_EXECUTION_FAILED:
return “NVJPEG_STATUS_EXECUTION_FAILED”;
case NVJPEG_STATUS_ARCH_MISMATCH:
return “NVJPEG_STATUS_ARCH_MISMATCH”;
case NVJPEG_STATUS_INTERNAL_ERROR:
return “NVJPEG_STATUS_INTERNAL_ERROR”;
return “
#ifdef NV_NPPIDEFS_H
// NPP API errors
static const char *_cudaGetErrorEnum(NppStatus error) {
switch (error) {
case NPP_NOT_SUPPORTED_MODE_ERROR:
return “NPP_NOT_SUPPORTED_MODE_ERROR”;
case NPP_ROUND_MODE_NOT_SUPPORTED_ERROR:
return “NPP_ROUND_MODE_NOT_SUPPORTED_ERROR”;
case NPP_RESIZE_NO_OPERATION_ERROR:
return “NPP_RESIZE_NO_OPERATION_ERROR”;
case NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY:
return “NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY”;
#if ((NPP_VERSION_MAJOR << 12) + (NPP_VERSION_MINOR << 4)) <= 0x5000
case NPP_BAD_ARG_ERROR:
return "NPP_BAD_ARGUMENT_ERROR";
case NPP_COEFF_ERROR:
return "NPP_COEFFICIENT_ERROR";
case NPP_RECT_ERROR:
return "NPP_RECTANGLE_ERROR";
case NPP_QUAD_ERROR:
return "NPP_QUADRANGLE_ERROR";
case NPP_MEM_ALLOC_ERR:
return "NPP_MEMORY_ALLOCATION_ERROR";
case NPP_HISTO_NUMBER_OF_LEVELS_ERROR:
return "NPP_HISTOGRAM_NUMBER_OF_LEVELS_ERROR";
case NPP_INVALID_INPUT:
return "NPP_INVALID_INPUT";
case NPP_POINTER_ERROR:
return "NPP_POINTER_ERROR";
case NPP_WARNING:
return "NPP_WARNING";
case NPP_ODD_ROI_WARNING:
return "NPP_ODD_ROI_WARNING";
// These are for CUDA 5.5 or higher
case NPP_BAD_ARGUMENT_ERROR:
return "NPP_BAD_ARGUMENT_ERROR";
case NPP_COEFFICIENT_ERROR:
return "NPP_COEFFICIENT_ERROR";
case NPP_RECTANGLE_ERROR:
return "NPP_RECTANGLE_ERROR";
case NPP_QUADRANGLE_ERROR:
return "NPP_QUADRANGLE_ERROR";
case NPP_MEMORY_ALLOCATION_ERR:
return "NPP_MEMORY_ALLOCATION_ERROR";
case NPP_HISTOGRAM_NUMBER_OF_LEVELS_ERROR:
return "NPP_HISTOGRAM_NUMBER_OF_LEVELS_ERROR";
case NPP_INVALID_HOST_POINTER_ERROR:
return "NPP_INVALID_HOST_POINTER_ERROR";
case NPP_INVALID_DEVICE_POINTER_ERROR:
return "NPP_INVALID_DEVICE_POINTER_ERROR";
case NPP_LUT_NUMBER_OF_LEVELS_ERROR:
return "NPP_LUT_NUMBER_OF_LEVELS_ERROR";
case NPP_TEXTURE_BIND_ERROR:
return "NPP_TEXTURE_BIND_ERROR";
case NPP_WRONG_INTERSECTION_ROI_ERROR:
return "NPP_WRONG_INTERSECTION_ROI_ERROR";
case NPP_NOT_EVEN_STEP_ERROR:
return "NPP_NOT_EVEN_STEP_ERROR";
case NPP_INTERPOLATION_ERROR:
return "NPP_INTERPOLATION_ERROR";
case NPP_RESIZE_FACTOR_ERROR:
return "NPP_RESIZE_FACTOR_ERROR";
case NPP_HAAR_CLASSIFIER_PIXEL_MATCH_ERROR:
return "NPP_HAAR_CLASSIFIER_PIXEL_MATCH_ERROR";
#if ((NPP_VERSION_MAJOR << 12) + (NPP_VERSION_MINOR << 4)) <= 0x5000
case NPP_MEMFREE_ERR:
return "NPP_MEMFREE_ERR";
case NPP_MEMSET_ERR:
return "NPP_MEMSET_ERR";
case NPP_MEMCPY_ERR:
return "NPP_MEMCPY_ERROR";
case NPP_MIRROR_FLIP_ERR:
return "NPP_MIRROR_FLIP_ERR";
case NPP_MEMFREE_ERROR:
return "NPP_MEMFREE_ERROR";
case NPP_MEMSET_ERROR:
return "NPP_MEMSET_ERROR";
case NPP_MEMCPY_ERROR:
return "NPP_MEMCPY_ERROR";
case NPP_MIRROR_FLIP_ERROR:
return "NPP_MIRROR_FLIP_ERROR";
case NPP_ALIGNMENT_ERROR:
return "NPP_ALIGNMENT_ERROR";
case NPP_STEP_ERROR:
return "NPP_STEP_ERROR";
case NPP_SIZE_ERROR:
return "NPP_SIZE_ERROR";
case NPP_NULL_POINTER_ERROR:
return "NPP_NULL_POINTER_ERROR";
case NPP_CUDA_KERNEL_EXECUTION_ERROR:
return "NPP_CUDA_KERNEL_EXECUTION_ERROR";
case NPP_NOT_IMPLEMENTED_ERROR:
return "NPP_NOT_IMPLEMENTED_ERROR";
case NPP_ERROR:
return "NPP_ERROR";
case NPP_SUCCESS:
return "NPP_SUCCESS";
case NPP_WRONG_INTERSECTION_QUAD_WARNING:
return "NPP_WRONG_INTERSECTION_QUAD_WARNING";
case NPP_MISALIGNED_DST_ROI_WARNING:
return "NPP_MISALIGNED_DST_ROI_WARNING";
case NPP_AFFINE_QUAD_INCORRECT_WARNING:
return "NPP_AFFINE_QUAD_INCORRECT_WARNING";
case NPP_DOUBLE_SIZE_WARNING:
return "NPP_DOUBLE_SIZE_WARNING";
case NPP_WRONG_INTERSECTION_ROI_WARNING:
return "NPP_WRONG_INTERSECTION_ROI_WARNING";
#if ((NPP_VERSION_MAJOR << 12) + (NPP_VERSION_MINOR << 4)) >= 0x6000
/* These are 6.0 or higher */
case NPP_LUT_PALETTE_BITSIZE_ERROR:
return “NPP_LUT_PALETTE_BITSIZE_ERROR”;
case NPP_ZC_MODE_NOT_SUPPORTED_ERROR:
return “NPP_ZC_MODE_NOT_SUPPORTED_ERROR”;
case NPP_QUALITY_INDEX_ERROR:
return “NPP_QUALITY_INDEX_ERROR”;
case NPP_CHANNEL_ORDER_ERROR:
return “NPP_CHANNEL_ORDER_ERROR”;
case NPP_ZERO_MASK_VALUE_ERROR:
return “NPP_ZERO_MASK_VALUE_ERROR”;
case NPP_NUMBER_OF_CHANNELS_ERROR:
return “NPP_NUMBER_OF_CHANNELS_ERROR”;
case NPP_COI_ERROR:
return “NPP_COI_ERROR”;
case NPP_DIVISOR_ERROR:
return “NPP_DIVISOR_ERROR”;
case NPP_CHANNEL_ERROR:
return “NPP_CHANNEL_ERROR”;
case NPP_STRIDE_ERROR:
return “NPP_STRIDE_ERROR”;
case NPP_ANCHOR_ERROR:
return “NPP_ANCHOR_ERROR”;
case NPP_MASK_SIZE_ERROR:
return “NPP_MASK_SIZE_ERROR”;
case NPP_MOMENT_00_ZERO_ERROR:
return “NPP_MOMENT_00_ZERO_ERROR”;
case NPP_THRESHOLD_NEGATIVE_LEVEL_ERROR:
return “NPP_THRESHOLD_NEGATIVE_LEVEL_ERROR”;
case NPP_THRESHOLD_ERROR:
return “NPP_THRESHOLD_ERROR”;
case NPP_CONTEXT_MATCH_ERROR:
return “NPP_CONTEXT_MATCH_ERROR”;
case NPP_FFT_FLAG_ERROR:
return “NPP_FFT_FLAG_ERROR”;
case NPP_FFT_ORDER_ERROR:
return “NPP_FFT_ORDER_ERROR”;
case NPP_SCALE_RANGE_ERROR:
return “NPP_SCALE_RANGE_ERROR”;
case NPP_DATA_TYPE_ERROR:
return “NPP_DATA_TYPE_ERROR”;
case NPP_OUT_OFF_RANGE_ERROR:
return “NPP_OUT_OFF_RANGE_ERROR”;
case NPP_DIVIDE_BY_ZERO_ERROR:
return “NPP_DIVIDE_BY_ZERO_ERROR”;
case NPP_RANGE_ERROR:
return “NPP_RANGE_ERROR”;
case NPP_NO_MEMORY_ERROR:
return “NPP_NO_MEMORY_ERROR”;
case NPP_ERROR_RESERVED:
return “NPP_ERROR_RESERVED”;
case NPP_NO_OPERATION_WARNING:
return “NPP_NO_OPERATION_WARNING”;
case NPP_DIVIDE_BY_ZERO_WARNING:
return “NPP_DIVIDE_BY_ZERO_WARNING”;
#if ((NPP_VERSION_MAJOR << 12) + (NPP_VERSION_MINOR << 4)) >= 0x7000
/* These are 7.0 or higher */
case NPP_OVERFLOW_ERROR:
return “NPP_OVERFLOW_ERROR”;
case NPP_CORRUPTED_DATA_ERROR:
return “NPP_CORRUPTED_DATA_ERROR”;
return “
template
void check(T result, char const *const func, const char *const file,
int const line) {
if (result) {
fprintf(stderr, “CUDA error at %s:%d code=%d(%s) \”%s\” \n”, file, line,
static_cast
exit(EXIT_FAILURE);
#ifdef __DRIVER_TYPES_H__
// This will output the proper CUDA error strings in the event
// that a CUDA host call returns an error
#define checkCudaErrors(val) check((val), #val, __FILE__, __LINE__)
// This will output the proper error string when calling cudaGetLastError
#define getLastCudaError(msg) __getLastCudaError(msg, __FILE__, __LINE__)
inline void __getLastCudaError(const char *errorMessage, const char *file,
const int line) {
cudaError_t err = cudaGetLastError();
if (cudaSuccess != err) {
fprintf(stderr,
“%s(%i) : getLastCudaError() CUDA error :”
” %s : (%d) %s.\n”,
file, line, errorMessage, static_cast
cudaGetErrorString(err));
exit(EXIT_FAILURE);
// This will only print the proper error string when calling cudaGetLastError
// but not exit program incase error detected.
#define printLastCudaError(msg) __printLastCudaError(msg, __FILE__, __LINE__)
inline void __printLastCudaError(const char *errorMessage, const char *file,
const int line) {
cudaError_t err = cudaGetLastError();
if (cudaSuccess != err) {
fprintf(stderr,
“%s(%i) : getLastCudaError() CUDA error :”
” %s : (%d) %s.\n”,
file, line, errorMessage, static_cast
cudaGetErrorString(err));
#ifndef MAX
#define MAX(a, b) (a > b ? a : b)
// Float To Int conversion
inline int ftoi(float value) {
return (value >= 0 ? static_cast
: static_cast
// Beginning of GPU Architecture definitions
inline int _ConvertSMVer2Cores(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},
{0x32, 192},
{0x35, 192},
{0x37, 192},
{0x50, 128},
{0x52, 128},
{0x53, 128},
{0x60, 64},
{0x61, 128},
{0x62, 128},
{0x70, 64},
{0x72, 64},
{0x75, 64},
{0x80, 64},
{0x86, 128},
{-1, -1}};
int index = 0;
while (nGpuArchCoresPerSM[index].SM != -1) {
if (nGpuArchCoresPerSM[index].SM == ((major << 4) + minor)) {
return nGpuArchCoresPerSM[index].Cores;
// If we don't find the values, we default use the previous one
// to run properly
"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;
inline const char* _ConvertSMVer2ArchName(int major, int minor) {
// Defines for GPU Architecture types (using the SM version to determine
// the GPU Arch name)
typedef struct {
int SM; // 0xMm (hexidecimal notation), M = SM Major version,
// and m = SM minor version
const char* name;
} sSMtoArchName;
sSMtoArchName nGpuArchNameSM[] = {
{0x30, "Kepler"},
{0x32, "Kepler"},
{0x35, "Kepler"},
{0x37, "Kepler"},
{0x50, "Maxwell"},
{0x52, "Maxwell"},
{0x53, "Maxwell"},
{0x60, "Pascal"},
{0x61, "Pascal"},
{0x62, "Pascal"},
{0x70, "Volta"},
{0x72, "Xavier"},
{0x75, "Turing"},
{0x80, "Ampere"},
{0x86, "Ampere"},
{-1, "Graphics Device"}};
int index = 0;
while (nGpuArchNameSM[index].SM != -1) {
if (nGpuArchNameSM[index].SM == ((major << 4) + minor)) {
return nGpuArchNameSM[index].name;
// If we don't find the values, we default use the previous one
// to run properly
"MapSMtoArchName for SM %d.%d is undefined."
" Default to use %s\n",
major, minor, nGpuArchNameSM[index - 1].name);
return nGpuArchNameSM[index - 1].name;
// end of GPU Architecture definitions
#ifdef __CUDA_RUNTIME_H__
// General GPU Device CUDA Initialization
inline int gpuDeviceInit(int devID) {
int device_count;
checkCudaErrors(cudaGetDeviceCount(&device_count));
if (device_count == 0) {
fprintf(stderr,
"gpuDeviceInit() CUDA error: "
"no devices supporting CUDA.\n");
exit(EXIT_FAILURE);
if (devID < 0) {
devID = 0;
if (devID > device_count – 1) {
fprintf(stderr, “\n”);
fprintf(stderr, “>> %d CUDA capable GPU device(s) detected. <<\n",
device_count);
fprintf(stderr,
">> gpuDeviceInit (-device=%d) is not a valid”
” GPU device. <<\n",
fprintf(stderr, "\n");
return -devID;
int computeMode = -1, major = 0, minor = 0;
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, devID));
checkCudaErrors(cudaDeviceGetAttribute(&major, cudaDevAttrComputeCapabilityMajor, devID));
checkCudaErrors(cudaDeviceGetAttribute(&minor, cudaDevAttrComputeCapabilityMinor, devID));
if (computeMode == cudaComputeModeProhibited) {
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com