mirror of
https://github.com/NVIDIA/cuda-samples.git
synced 2026-10-26 06:38:16 +08:00
Apply consistent code formatting across the repo. Add clang-format and pre-commit hooks.
This commit is contained in:
Rob Armstrong
@@ -36,215 +36,225 @@
|
||||
#include <ImageIO.h>
|
||||
#include <ImagesCPU.h>
|
||||
#include <ImagesNPP.h>
|
||||
|
||||
#include <string.h>
|
||||
#include <fstream>
|
||||
#include <iostream>
|
||||
|
||||
#include <cuda_runtime.h>
|
||||
#include <npp.h>
|
||||
|
||||
#include <fstream>
|
||||
#include <helper_cuda.h>
|
||||
#include <helper_string.h>
|
||||
#include <iostream>
|
||||
#include <npp.h>
|
||||
#include <string.h>
|
||||
|
||||
inline int cudaDeviceInit(int argc, const char **argv) {
|
||||
int deviceCount;
|
||||
checkCudaErrors(cudaGetDeviceCount(&deviceCount));
|
||||
inline int cudaDeviceInit(int argc, const char **argv)
|
||||
{
|
||||
int deviceCount;
|
||||
checkCudaErrors(cudaGetDeviceCount(&deviceCount));
|
||||
|
||||
if (deviceCount == 0) {
|
||||
std::cerr << "CUDA error: no devices supporting CUDA." << std::endl;
|
||||
exit(EXIT_FAILURE);
|
||||
}
|
||||
if (deviceCount == 0) {
|
||||
std::cerr << "CUDA error: no devices supporting CUDA." << std::endl;
|
||||
exit(EXIT_FAILURE);
|
||||
}
|
||||
|
||||
int dev = findCudaDevice(argc, argv);
|
||||
int dev = findCudaDevice(argc, argv);
|
||||
|
||||
cudaDeviceProp deviceProp;
|
||||
cudaGetDeviceProperties(&deviceProp, dev);
|
||||
std::cerr << "cudaSetDevice GPU" << dev << " = " << deviceProp.name
|
||||
<< std::endl;
|
||||
cudaDeviceProp deviceProp;
|
||||
cudaGetDeviceProperties(&deviceProp, dev);
|
||||
std::cerr << "cudaSetDevice GPU" << dev << " = " << deviceProp.name << std::endl;
|
||||
|
||||
checkCudaErrors(cudaSetDevice(dev));
|
||||
checkCudaErrors(cudaSetDevice(dev));
|
||||
|
||||
return dev;
|
||||
return dev;
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
printf("%s Starting...\n\n", argv[0]);
|
||||
int main(int argc, char *argv[])
|
||||
{
|
||||
printf("%s Starting...\n\n", argv[0]);
|
||||
|
||||
try {
|
||||
std::string sFilename;
|
||||
char *filePath;
|
||||
try {
|
||||
std::string sFilename;
|
||||
char *filePath;
|
||||
|
||||
cudaDeviceInit(argc, (const char **)argv);
|
||||
cudaDeviceInit(argc, (const char **)argv);
|
||||
|
||||
NppStreamContext nppStreamCtx;
|
||||
nppStreamCtx.hStream = 0; // The NULL stream by default, set this to whatever your stream ID is if not the NULL stream.
|
||||
NppStreamContext nppStreamCtx;
|
||||
nppStreamCtx.hStream =
|
||||
0; // The NULL stream by default, set this to whatever your stream ID is if not the NULL stream.
|
||||
|
||||
cudaError_t cudaError = cudaGetDevice(&nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess)
|
||||
{
|
||||
printf("CUDA error: no devices supporting CUDA.\n");
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
cudaError_t cudaError = cudaGetDevice(&nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess) {
|
||||
printf("CUDA error: no devices supporting CUDA.\n");
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
}
|
||||
|
||||
const NppLibraryVersion *libVer = nppGetLibVersion();
|
||||
|
||||
printf("NPP Library Version %d.%d.%d\n", libVer->major, libVer->minor, libVer->build);
|
||||
|
||||
int driverVersion, runtimeVersion;
|
||||
cudaDriverGetVersion(&driverVersion);
|
||||
cudaRuntimeGetVersion(&runtimeVersion);
|
||||
|
||||
printf("CUDA Driver Version: %d.%d\n", driverVersion / 1000, (driverVersion % 100) / 10);
|
||||
printf("CUDA Runtime Version: %d.%d\n\n", runtimeVersion / 1000, (runtimeVersion % 100) / 10);
|
||||
|
||||
cudaError = cudaDeviceGetAttribute(&nppStreamCtx.nCudaDevAttrComputeCapabilityMajor,
|
||||
cudaDevAttrComputeCapabilityMajor,
|
||||
nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess)
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
|
||||
cudaError = cudaDeviceGetAttribute(&nppStreamCtx.nCudaDevAttrComputeCapabilityMinor,
|
||||
cudaDevAttrComputeCapabilityMinor,
|
||||
nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess)
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
|
||||
cudaError = cudaStreamGetFlags(nppStreamCtx.hStream, &nppStreamCtx.nStreamFlags);
|
||||
|
||||
cudaDeviceProp oDeviceProperties;
|
||||
|
||||
cudaError = cudaGetDeviceProperties(&oDeviceProperties, nppStreamCtx.nCudaDeviceId);
|
||||
|
||||
nppStreamCtx.nMultiProcessorCount = oDeviceProperties.multiProcessorCount;
|
||||
nppStreamCtx.nMaxThreadsPerMultiProcessor = oDeviceProperties.maxThreadsPerMultiProcessor;
|
||||
nppStreamCtx.nMaxThreadsPerBlock = oDeviceProperties.maxThreadsPerBlock;
|
||||
nppStreamCtx.nSharedMemPerBlock = oDeviceProperties.sharedMemPerBlock;
|
||||
|
||||
if (checkCmdLineFlag(argc, (const char **)argv, "input")) {
|
||||
getCmdLineArgumentString(argc, (const char **)argv, "input", &filePath);
|
||||
}
|
||||
else {
|
||||
filePath = sdkFindFilePath("teapot512.pgm", argv[0]);
|
||||
}
|
||||
|
||||
if (filePath) {
|
||||
sFilename = filePath;
|
||||
}
|
||||
else {
|
||||
sFilename = "teapot512.pgm";
|
||||
}
|
||||
|
||||
// if we specify the filename at the command line, then we only test
|
||||
// sFilename[0].
|
||||
int file_errors = 0;
|
||||
std::ifstream infile(sFilename.data(), std::ifstream::in);
|
||||
|
||||
if (infile.good()) {
|
||||
std::cout << "cannyEdgeDetectionNPP opened: <" << sFilename.data() << "> successfully!" << std::endl;
|
||||
file_errors = 0;
|
||||
infile.close();
|
||||
}
|
||||
else {
|
||||
std::cout << "cannyEdgeDetectionNPP unable to open: <" << sFilename.data() << ">" << std::endl;
|
||||
file_errors++;
|
||||
infile.close();
|
||||
}
|
||||
|
||||
if (file_errors > 0) {
|
||||
exit(EXIT_FAILURE);
|
||||
}
|
||||
|
||||
std::string sResultFilename = sFilename;
|
||||
|
||||
std::string::size_type dot = sResultFilename.rfind('.');
|
||||
|
||||
if (dot != std::string::npos) {
|
||||
sResultFilename = sResultFilename.substr(0, dot);
|
||||
}
|
||||
|
||||
sResultFilename += "_cannyEdgeDetection.pgm";
|
||||
|
||||
if (checkCmdLineFlag(argc, (const char **)argv, "output")) {
|
||||
char *outputFilePath;
|
||||
getCmdLineArgumentString(argc, (const char **)argv, "output", &outputFilePath);
|
||||
sResultFilename = outputFilePath;
|
||||
}
|
||||
|
||||
// declare a host image object for an 8-bit grayscale image
|
||||
npp::ImageCPU_8u_C1 oHostSrc;
|
||||
// load gray-scale image from disk
|
||||
npp::loadImage(sFilename, oHostSrc);
|
||||
// declare a device image and copy construct from the host image,
|
||||
// i.e. upload host to device
|
||||
npp::ImageNPP_8u_C1 oDeviceSrc(oHostSrc);
|
||||
|
||||
NppiSize oSrcSize = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
|
||||
NppiPoint oSrcOffset = {0, 0};
|
||||
|
||||
// create struct with ROI size
|
||||
NppiSize oSizeROI = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
|
||||
// allocate device image of appropriately reduced size
|
||||
npp::ImageNPP_8u_C1 oDeviceDst(oSizeROI.width, oSizeROI.height);
|
||||
|
||||
int nBufferSize = 0;
|
||||
Npp8u *pScratchBufferNPP = 0;
|
||||
|
||||
// get necessary scratch buffer size and allocate that much device memory
|
||||
NPP_CHECK_NPP(nppiFilterCannyBorderGetBufferSize(oSizeROI, &nBufferSize));
|
||||
|
||||
cudaMalloc((void **)&pScratchBufferNPP, nBufferSize);
|
||||
|
||||
// now run the canny edge detection filter
|
||||
// Using nppiNormL2 will produce larger magnitude values allowing for finer
|
||||
// control of threshold values while nppiNormL1 will be slightly faster.
|
||||
// Also, selecting the sobel gradient filter allows up to a 5x5 kernel size
|
||||
// which can produce more precise results but is a bit slower. Commonly
|
||||
// nppiNormL2 and sobel gradient filter size of 3x3 are used. Canny
|
||||
// recommends that the high threshold value should be about 3 times the low
|
||||
// threshold value. The threshold range will depend on the range of
|
||||
// magnitude values that the sobel gradient filter generates for a
|
||||
// particular image.
|
||||
|
||||
Npp16s nLowThreshold = 72;
|
||||
Npp16s nHighThreshold = 256;
|
||||
|
||||
if ((nBufferSize > 0) && (pScratchBufferNPP != 0)) {
|
||||
NPP_CHECK_NPP(nppiFilterCannyBorder_8u_C1R_Ctx(oDeviceSrc.data(),
|
||||
oDeviceSrc.pitch(),
|
||||
oSrcSize,
|
||||
oSrcOffset,
|
||||
oDeviceDst.data(),
|
||||
oDeviceDst.pitch(),
|
||||
oSizeROI,
|
||||
NPP_FILTER_SOBEL,
|
||||
NPP_MASK_SIZE_3_X_3,
|
||||
nLowThreshold,
|
||||
nHighThreshold,
|
||||
nppiNormL2,
|
||||
NPP_BORDER_REPLICATE,
|
||||
pScratchBufferNPP,
|
||||
nppStreamCtx));
|
||||
}
|
||||
|
||||
// free scratch buffer memory
|
||||
cudaFree(pScratchBufferNPP);
|
||||
|
||||
// declare a host image for the result
|
||||
npp::ImageCPU_8u_C1 oHostDst(oDeviceDst.size());
|
||||
// and copy the device result data into it
|
||||
oDeviceDst.copyTo(oHostDst.data(), oHostDst.pitch());
|
||||
|
||||
saveImage(sResultFilename, oHostDst);
|
||||
std::cout << "Saved image: " << sResultFilename << std::endl;
|
||||
|
||||
nppiFree(oDeviceSrc.data());
|
||||
nppiFree(oDeviceDst.data());
|
||||
|
||||
exit(EXIT_SUCCESS);
|
||||
}
|
||||
catch (npp::Exception &rException) {
|
||||
std::cerr << "Program error! The following exception occurred: \n";
|
||||
std::cerr << rException << std::endl;
|
||||
std::cerr << "Aborting." << std::endl;
|
||||
|
||||
exit(EXIT_FAILURE);
|
||||
}
|
||||
catch (...) {
|
||||
std::cerr << "Program error! An unknow type of exception occurred. \n";
|
||||
std::cerr << "Aborting." << std::endl;
|
||||
|
||||
exit(EXIT_FAILURE);
|
||||
return -1;
|
||||
}
|
||||
|
||||
const NppLibraryVersion *libVer = nppGetLibVersion();
|
||||
|
||||
printf("NPP Library Version %d.%d.%d\n", libVer->major, libVer->minor, libVer->build);
|
||||
|
||||
int driverVersion, runtimeVersion;
|
||||
cudaDriverGetVersion(&driverVersion);
|
||||
cudaRuntimeGetVersion(&runtimeVersion);
|
||||
|
||||
printf("CUDA Driver Version: %d.%d\n", driverVersion/1000, (driverVersion%100)/10);
|
||||
printf("CUDA Runtime Version: %d.%d\n\n", runtimeVersion/1000, (runtimeVersion%100)/10);
|
||||
|
||||
cudaError = cudaDeviceGetAttribute(&nppStreamCtx.nCudaDevAttrComputeCapabilityMajor,
|
||||
cudaDevAttrComputeCapabilityMajor,
|
||||
nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess)
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
|
||||
cudaError = cudaDeviceGetAttribute(&nppStreamCtx.nCudaDevAttrComputeCapabilityMinor,
|
||||
cudaDevAttrComputeCapabilityMinor,
|
||||
nppStreamCtx.nCudaDeviceId);
|
||||
if (cudaError != cudaSuccess)
|
||||
return NPP_NOT_SUFFICIENT_COMPUTE_CAPABILITY;
|
||||
|
||||
cudaError = cudaStreamGetFlags(nppStreamCtx.hStream, &nppStreamCtx.nStreamFlags);
|
||||
|
||||
cudaDeviceProp oDeviceProperties;
|
||||
|
||||
cudaError = cudaGetDeviceProperties(&oDeviceProperties, nppStreamCtx.nCudaDeviceId);
|
||||
|
||||
nppStreamCtx.nMultiProcessorCount = oDeviceProperties.multiProcessorCount;
|
||||
nppStreamCtx.nMaxThreadsPerMultiProcessor = oDeviceProperties.maxThreadsPerMultiProcessor;
|
||||
nppStreamCtx.nMaxThreadsPerBlock = oDeviceProperties.maxThreadsPerBlock;
|
||||
nppStreamCtx.nSharedMemPerBlock = oDeviceProperties.sharedMemPerBlock;
|
||||
|
||||
if (checkCmdLineFlag(argc, (const char **)argv, "input")) {
|
||||
getCmdLineArgumentString(argc, (const char **)argv, "input", &filePath);
|
||||
} else {
|
||||
filePath = sdkFindFilePath("teapot512.pgm", argv[0]);
|
||||
}
|
||||
|
||||
if (filePath) {
|
||||
sFilename = filePath;
|
||||
} else {
|
||||
sFilename = "teapot512.pgm";
|
||||
}
|
||||
|
||||
// if we specify the filename at the command line, then we only test
|
||||
// sFilename[0].
|
||||
int file_errors = 0;
|
||||
std::ifstream infile(sFilename.data(), std::ifstream::in);
|
||||
|
||||
if (infile.good()) {
|
||||
std::cout << "cannyEdgeDetectionNPP opened: <" << sFilename.data()
|
||||
<< "> successfully!" << std::endl;
|
||||
file_errors = 0;
|
||||
infile.close();
|
||||
} else {
|
||||
std::cout << "cannyEdgeDetectionNPP unable to open: <" << sFilename.data()
|
||||
<< ">" << std::endl;
|
||||
file_errors++;
|
||||
infile.close();
|
||||
}
|
||||
|
||||
if (file_errors > 0) {
|
||||
exit(EXIT_FAILURE);
|
||||
}
|
||||
|
||||
std::string sResultFilename = sFilename;
|
||||
|
||||
std::string::size_type dot = sResultFilename.rfind('.');
|
||||
|
||||
if (dot != std::string::npos) {
|
||||
sResultFilename = sResultFilename.substr(0, dot);
|
||||
}
|
||||
|
||||
sResultFilename += "_cannyEdgeDetection.pgm";
|
||||
|
||||
if (checkCmdLineFlag(argc, (const char **)argv, "output")) {
|
||||
char *outputFilePath;
|
||||
getCmdLineArgumentString(argc, (const char **)argv, "output",
|
||||
&outputFilePath);
|
||||
sResultFilename = outputFilePath;
|
||||
}
|
||||
|
||||
// declare a host image object for an 8-bit grayscale image
|
||||
npp::ImageCPU_8u_C1 oHostSrc;
|
||||
// load gray-scale image from disk
|
||||
npp::loadImage(sFilename, oHostSrc);
|
||||
// declare a device image and copy construct from the host image,
|
||||
// i.e. upload host to device
|
||||
npp::ImageNPP_8u_C1 oDeviceSrc(oHostSrc);
|
||||
|
||||
NppiSize oSrcSize = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
|
||||
NppiPoint oSrcOffset = {0, 0};
|
||||
|
||||
// create struct with ROI size
|
||||
NppiSize oSizeROI = {(int)oDeviceSrc.width(), (int)oDeviceSrc.height()};
|
||||
// allocate device image of appropriately reduced size
|
||||
npp::ImageNPP_8u_C1 oDeviceDst(oSizeROI.width, oSizeROI.height);
|
||||
|
||||
int nBufferSize = 0;
|
||||
Npp8u *pScratchBufferNPP = 0;
|
||||
|
||||
// get necessary scratch buffer size and allocate that much device memory
|
||||
NPP_CHECK_NPP(nppiFilterCannyBorderGetBufferSize(oSizeROI, &nBufferSize));
|
||||
|
||||
cudaMalloc((void **)&pScratchBufferNPP, nBufferSize);
|
||||
|
||||
// now run the canny edge detection filter
|
||||
// Using nppiNormL2 will produce larger magnitude values allowing for finer
|
||||
// control of threshold values while nppiNormL1 will be slightly faster.
|
||||
// Also, selecting the sobel gradient filter allows up to a 5x5 kernel size
|
||||
// which can produce more precise results but is a bit slower. Commonly
|
||||
// nppiNormL2 and sobel gradient filter size of 3x3 are used. Canny
|
||||
// recommends that the high threshold value should be about 3 times the low
|
||||
// threshold value. The threshold range will depend on the range of
|
||||
// magnitude values that the sobel gradient filter generates for a
|
||||
// particular image.
|
||||
|
||||
Npp16s nLowThreshold = 72;
|
||||
Npp16s nHighThreshold = 256;
|
||||
|
||||
if ((nBufferSize > 0) && (pScratchBufferNPP != 0)) {
|
||||
NPP_CHECK_NPP(nppiFilterCannyBorder_8u_C1R_Ctx(
|
||||
oDeviceSrc.data(), oDeviceSrc.pitch(), oSrcSize, oSrcOffset,
|
||||
oDeviceDst.data(), oDeviceDst.pitch(), oSizeROI, NPP_FILTER_SOBEL,
|
||||
NPP_MASK_SIZE_3_X_3, nLowThreshold, nHighThreshold, nppiNormL2,
|
||||
NPP_BORDER_REPLICATE, pScratchBufferNPP, nppStreamCtx));
|
||||
}
|
||||
|
||||
// free scratch buffer memory
|
||||
cudaFree(pScratchBufferNPP);
|
||||
|
||||
// declare a host image for the result
|
||||
npp::ImageCPU_8u_C1 oHostDst(oDeviceDst.size());
|
||||
// and copy the device result data into it
|
||||
oDeviceDst.copyTo(oHostDst.data(), oHostDst.pitch());
|
||||
|
||||
saveImage(sResultFilename, oHostDst);
|
||||
std::cout << "Saved image: " << sResultFilename << std::endl;
|
||||
|
||||
nppiFree(oDeviceSrc.data());
|
||||
nppiFree(oDeviceDst.data());
|
||||
|
||||
exit(EXIT_SUCCESS);
|
||||
} catch (npp::Exception &rException) {
|
||||
std::cerr << "Program error! The following exception occurred: \n";
|
||||
std::cerr << rException << std::endl;
|
||||
std::cerr << "Aborting." << std::endl;
|
||||
|
||||
exit(EXIT_FAILURE);
|
||||
} catch (...) {
|
||||
std::cerr << "Program error! An unknow type of exception occurred. \n";
|
||||
std::cerr << "Aborting." << std::endl;
|
||||
|
||||
exit(EXIT_FAILURE);
|
||||
return -1;
|
||||
}
|
||||
|
||||
return 0;
|
||||
return 0;
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user