Skip to content

Dev0 #14

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
wffpy merged 2 commits into from
Dec 10, 2024
Merged

Dev0 #14

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5593f0d
hook cuLaunchKernelEx
wffpy Dec 10, 2024
d0e26fe
fix gpu hook cuLaunchKernelEx
wffpy Dec 10, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,8 +22,14 @@ else ()
message(STATUS "CUDA is not enabled")
endif()

if (USE_CAPSTONE STREQUAL TRUE)
ADD_DEFINITIONS(-DUSE_CAPSTONE)
message("CAPSTONE is enabled!!!!!!!!!!")
endif()

if (XPU_DEV)
message("XPU_DEB: ${XPU_DEV}")
ADD_DEFINITIONS(-DXPU_DEV)
find_library(XPU_LIB xpurt PATHS /usr/local/xcudart/lib)
message("XPU_LIB: ${XPU_LIB}")
endif()
Expand Down
14 changes: 14 additions & 0 deletions lib/cuda/GpuProfiler.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -320,6 +320,7 @@ CUresult local_cuLaunchKernel(CUfunction f,
void **kernelParams,
void **extra) {
trace::Tracer tracer(__FUNCTION__);
LOG() << "call !!!!!!!";
cudaDeviceSynchronize();
CUresult ret = Target_cuLaunchKernel(f,
gridDimX,
Expand All @@ -337,8 +338,21 @@ CUresult local_cuLaunchKernel(CUfunction f,
return ret;
}


CUresult (*Target_cuLaunchKernelEx) ( const CUlaunchConfig* config, CUfunction f, void** kernelParams, void** extra ) = nullptr;
CUresult local_cuLaunchKernelEx ( const CUlaunchConfig* config, CUfunction f, void** kernelParams, void** extra ) {
LOG() << "call !!!!!!!";
trace::Tracer tracer(__FUNCTION__);
cudaDeviceSynchronize();
CUresult ret = Target_cuLaunchKernelEx(config, f, kernelParams, extra);
cudaDeviceSynchronize();
cupti_activity_flush();
return ret;
}

// REGISTER_LOCAL_HOOK(cuGetProcAddress_v2, (void*)local_cuGetProcAddress, (void**)&Target_cuGetProcAddress);
REGISTER_LOCAL_HOOK(cuLaunchKernel, (void*)local_cuLaunchKernel, (void**)&Target_cuLaunchKernel);
REGISTER_LOCAL_HOOK(cuLaunchKernelEx, (void*)local_cuLaunchKernelEx, (void**)&Target_cuLaunchKernelEx);
#else
// Non-Gpu device
namespace gpu_profiler {
Expand Down