potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-30 15:53:03 +04:00
Code Packages Releases Wiki Activity

Merge remote-tracking branch 'upstream/2.4' into merge-2.4

Browse Source 
* #1538 from StevenPuttemans:bugfix_3283
* #1545 from alalek:ocl_test_fix_rng
* #1551 from alalek:cmake_install_win
* #1570 from ilya-lavrenov:ipp_warn_fix
* #1573 from alalek:perf_simple_strategy
* #1574 from alalek:svm_workaround
* #1576 from alalek:ocl_fix_cl_double
* #1577 from ilya-lavrenov:ocl_setto_opencl12
* #1578 from asmorkalov:android_fd_cp_fix
* #1579 from ilya-lavrenov:ocl_norm
* #1582 from sperrholz:ocl-arithm-additions
* #1586 from ilya-lavrenov:ocl_setto_win_fix
* #1589 from ilya-lavrenov:pr1582_fix
* #1591 from alalek:ocl_remove_cl_hpp_h
* #1592 from alalek:ocl_program_cache_update
* #1593 from ilya-lavrenov:ocl_war_on_double
* #1594 from ilya-lavrenov:ocl_perf
* #1595 from alalek:cl_code_cleanup
* #1596 from alalek:test_fix_run_py
* #1597 from alalek:ocl_fix_cleanup
* #1598 from alalek:ocl_fix_build_mac
* #1599 from ilya-lavrenov:ocl_mac_kernel_warnings
* #1601 from ilya-lavrenov:ocl_fix_tvl1_and_sparse
* #1602 from alalek:ocl_test_dump_info
* #1603 from ilya-lavrenov:ocl_disable_svm_noblas
* #1605 from alalek:ocl_fixes
* #1606 from ilya-lavrenov:ocl_imgproc
* #1607 from ilya-lavrenov:ocl_fft_cleanup
* #1608 from alalek:fix_warn_upd_haar
* #1609 from ilya-lavrenov:ocl_some_optimization
* #1610 from alalek:ocl_fix_perf_kalman
* #1612 from alalek:ocl_fix_string_info
* #1614 from ilya-lavrenov:ocl_svm_misprint
* #1616 from ilya-lavrenov:ocl_cvtColor
* #1617 from ilya-lavrenov:ocl_info
* #1622 from a0byte:2.4
* #1625 from ilya-lavrenov:to_string

Conflicts:
	cmake/OpenCVConfig.cmake
	cmake/OpenCVDetectPython.cmake
	cmake/OpenCVGenConfig.cmake
	modules/core/CMakeLists.txt
	modules/nonfree/src/surf.ocl.cpp
	modules/ocl/include/opencv2/ocl/ocl.hpp
	modules/ocl/include/opencv2/ocl/private/util.hpp
	modules/ocl/perf/main.cpp
	modules/ocl/src/arithm.cpp
	modules/ocl/src/cl_operations.cpp
	modules/ocl/src/cl_programcache.cpp
	modules/ocl/src/color.cpp
	modules/ocl/src/fft.cpp
	modules/ocl/src/filtering.cpp
	modules/ocl/src/gemm.cpp
	modules/ocl/src/haar.cpp
	modules/ocl/src/imgproc.cpp
	modules/ocl/src/matrix_operations.cpp
	modules/ocl/src/pyrlk.cpp
	modules/ocl/src/split_merge.cpp
	modules/ocl/src/svm.cpp
	modules/ocl/test/main.cpp
	modules/ocl/test/test_fft.cpp
	modules/ocl/test/test_moments.cpp
	modules/ocl/test/test_objdetect.cpp
	modules/ocl/test/test_optflow.cpp
	modules/ocl/test/utility.hpp
	modules/python/CMakeLists.txt
	modules/ts/include/opencv2/ts.hpp
	modules/ts/src/ts_perf.cpp
	samples/android/face-detection/jni/DetectionBasedTracker_jni.cpp
This commit is contained in:
Alexander Alekhin
2013-10-15 18:43:37 +04:00
parent 3f8db9d708 98d55f34fa
commit e845184843
124 changed files with 3144 additions and 3195 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/ocl/src/cl_operations.cpp
+46 -60
View File
@@ -212,13 +212,35 @@ void openCLVerifyKernel(const Context *ctx, cl_kernel kernel, size_t *localThrea
static double total_execute_time = 0;
static double total_kernel_time = 0;
#endif
static std::string removeDuplicatedWhiteSpaces(const char * buildOptions)
{
if (buildOptions == NULL)
return "";
size_t length = strlen(buildOptions), didx = 0, sidx = 0;
while (sidx < length && buildOptions[sidx] == 0)
++sidx;
std::string opt;
opt.resize(length);
for ( ; sidx < length; ++sidx)
if (buildOptions[sidx] != ' ')
opt[didx++] = buildOptions[sidx];
else if ( !(didx > 0 && opt[didx - 1] == ' ') )
opt[didx++] = buildOptions[sidx];
return opt;
}
void openCLExecuteKernel_(Context *ctx, const cv::ocl::ProgramEntry* source, String kernelName, size_t globalThreads[3],
size_t localThreads[3], std::vector< std::pair<size_t, const void *> > &args, int channels,
int depth, const char *build_options)
{
//construct kernel name
//The rule is functionName_Cn_Dn, C represent Channels, D Represent DataType Depth, n represent an integer number
//for exmaple split_C2_D2, represent the split kernel with channels =2 and dataType Depth = 2(Data type is char)
//for example split_C2_D3, represent the split kernel with channels = 2 and dataType Depth = 3(Data type is short)
std::stringstream idxStr;
if(channels != -1)
idxStr << "_C" << channels;
@@ -227,7 +249,8 @@ void openCLExecuteKernel_(Context *ctx, const cv::ocl::ProgramEntry* source, Str
kernelName = kernelName + idxStr.str();
cl_kernel kernel;
kernel = openCLGetKernelFromSource(ctx, source, kernelName, build_options);
std::string fixedOptions = removeDuplicatedWhiteSpaces(build_options);
kernel = openCLGetKernelFromSource(ctx, source, kernelName, fixedOptions.c_str());
if ( localThreads != NULL)
{
@@ -302,28 +325,27 @@ void openCLExecuteKernel(Context *ctx, const cv::ocl::ProgramEntry* source, Stri
total_kernel_time = 0;
cout << "-------------------------------------" << endl;
cout << setiosflags(ios::left) << setw(15) << "excute time";
cout << setiosflags(ios::left) << setw(15) << "lauch time";
cout << setiosflags(ios::left) << setw(15) << "execute time";
cout << setiosflags(ios::left) << setw(15) << "launch time";
cout << setiosflags(ios::left) << setw(15) << "kernel time" << endl;
int i = 0;
for(i = 0; i < RUN_TIMES; i++)
openCLExecuteKernel_(ctx, source, kernelName, globalThreads, localThreads, args, channels, depth,
build_options);
cout << "average kernel excute time: " << total_execute_time / RUN_TIMES << endl; // "ms" << endl;
cout << "average kernel execute time: " << total_execute_time / RUN_TIMES << endl; // "ms" << endl;
cout << "average kernel total time: " << total_kernel_time / RUN_TIMES << endl; // "ms" << endl;
#endif
}
double openCLExecuteKernelInterop(Context *ctx, const cv::ocl::ProgramEntry* source, String kernelName,
void openCLExecuteKernelInterop(Context *ctx, const cv::ocl::ProgramSource& source, String kernelName,
size_t globalThreads[3], size_t localThreads[3],
std::vector< std::pair<size_t, const void *> > &args, int channels, int depth, const char *build_options,
bool finish, bool measureKernelTime, bool cleanUp)
std::vector< std::pair<size_t, const void *> > &args, int channels, int depth, const char *build_options)
{
//construct kernel name
//The rule is functionName_Cn_Dn, C represent Channels, D Represent DataType Depth, n represent an integer number
//for exmaple split_C2_D2, represent the split kernel with channels =2 and dataType Depth = 2(Data type is char)
//for example split_C2_D2, represent the split kernel with channels = 2 and dataType Depth = 2 (Data type is char)
std::stringstream idxStr;
if(channels != -1)
idxStr << "_C" << channels;
@@ -331,63 +353,27 @@ double openCLExecuteKernelInterop(Context *ctx, const cv::ocl::ProgramEntry* sou
idxStr << "_D" << depth;
kernelName = kernelName + idxStr.str();
cl_kernel kernel;
kernel = openCLGetKernelFromSource(ctx, source, kernelName, build_options);
std::string name = std::string("custom_") + source.name;
ProgramEntry program = { name.c_str(), source.programStr, source.programHash };
cl_kernel kernel = openCLGetKernelFromSource(ctx, &program, kernelName, build_options);
double kernelTime = 0.0;
if( globalThreads != NULL)
CV_Assert(globalThreads != NULL);
if ( localThreads != NULL)
{
if ( localThreads != NULL)
{
globalThreads[0] = divUp(globalThreads[0], localThreads[0]) * localThreads[0];
globalThreads[1] = divUp(globalThreads[1], localThreads[1]) * localThreads[1];
globalThreads[2] = divUp(globalThreads[2], localThreads[2]) * localThreads[2];
globalThreads[0] = roundUp(globalThreads[0], localThreads[0]);
globalThreads[1] = roundUp(globalThreads[1], localThreads[1]);
globalThreads[2] = roundUp(globalThreads[2], localThreads[2]);
//size_t blockSize = localThreads[0] * localThreads[1] * localThreads[2];
cv::ocl::openCLVerifyKernel(ctx, kernel, localThreads);
}
for(size_t i = 0; i < args.size(); i ++)
openCLSafeCall(clSetKernelArg(kernel, i, args[i].first, args[i].second));
if(measureKernelTime == false)
{
openCLSafeCall(clEnqueueNDRangeKernel(getClCommandQueue(ctx), kernel, 3, NULL, globalThreads,
localThreads, 0, NULL, NULL));
}
else
{
cl_event event = NULL;
openCLSafeCall(clEnqueueNDRangeKernel(getClCommandQueue(ctx), kernel, 3, NULL, globalThreads,
localThreads, 0, NULL, &event));
cl_ulong end_time, queue_time;
openCLSafeCall(clWaitForEvents(1, &event));
openCLSafeCall(clGetEventProfilingInfo(event, CL_PROFILING_COMMAND_END,
sizeof(cl_ulong), &end_time, 0));
openCLSafeCall(clGetEventProfilingInfo(event, CL_PROFILING_COMMAND_QUEUED,
sizeof(cl_ulong), &queue_time, 0));
kernelTime = (double)(end_time - queue_time) / (1000 * 1000);
clReleaseEvent(event);
}
cv::ocl::openCLVerifyKernel(ctx, kernel, localThreads);
}
for(size_t i = 0; i < args.size(); i ++)
openCLSafeCall(clSetKernelArg(kernel, i, args[i].first, args[i].second));
if(finish)
{
clFinish(getClCommandQueue(ctx));
}
openCLSafeCall(clEnqueueNDRangeKernel(getClCommandQueue(ctx), kernel, 3, NULL, globalThreads,
localThreads, 0, NULL, NULL));
if(cleanUp)
{
openCLSafeCall(clReleaseKernel(kernel));
}
return kernelTime;
clFinish(getClCommandQueue(ctx));
openCLSafeCall(clReleaseKernel(kernel));
}
cl_mem load_constant(cl_context context, cl_command_queue command_queue, const void *value,