Skip to content

cuFFT: synchronize the plan stream before cufftXtMemcpy in the multi-GPU samples #367

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

Open
kevmo314 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

cuFFT: synchronize the plan stream before cufftXtMemcpy in the multi-GPU samples #367

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
5 changes: 5 additions & 0 deletions cuFFT/1d_mgpu_c2c/1d_mgpu_c2c_example.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,6 +105,11 @@ void spmg(dim_t fft, int &batch_size, gpus_t gpus, cpudata_t &h_data_in, cpudata
// Execute the plan
CUFFT_CALL(cufftXtExecDescriptor(plan, indesc, indesc, CUFFT_FORWARD));

#if CUFFT_VERSION >= 10400
// The transform runs on the plan's stream; wait for it before reading back
CUDA_RT_CALL(cudaStreamSynchronize(stream));
#endif

// Copy output data to CPU
CUFFT_CALL(cufftXtMemcpy(plan, reinterpret_cast<void *>(h_data_out.data()),
reinterpret_cast<void *>(indesc), CUFFT_COPY_DEVICE_TO_HOST));
Expand Down
5 changes: 5 additions & 0 deletions cuFFT/3d_mgpu_c2c/3d_mgpu_c2c_example.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,6 +105,11 @@ void spmg(dim_t fft, gpus_t gpus, cpudata_t &h_data_in, cpudata_t &h_data_out,
// Execute the plan
CUFFT_CALL(cufftXtExecDescriptor(plan, indesc, indesc, CUFFT_FORWARD));

#if CUFFT_VERSION >= 10400
// The transform runs on the plan's stream; wait for it before reading back
CUDA_RT_CALL(cudaStreamSynchronize(stream));
#endif

// Copy output data to CPU
CUFFT_CALL(cufftXtMemcpy(plan, reinterpret_cast<void *>(h_data_out.data()),
reinterpret_cast<void *>(indesc), CUFFT_COPY_DEVICE_TO_HOST));
Expand Down
10 changes: 10 additions & 0 deletions cuFFT/3d_mgpu_r2c_c2r/3d_mgpu_r2c_c2r_example.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,13 +155,23 @@ void spmg(dim_t fft, gpus_t gpus, cpudata_t &h_data_in, cpudata_t &h_data_out,
// Execute the plan_r2c
CUFFT_CALL(cufftXtExecDescriptor(plan_r2c, indesc, indesc, CUFFT_FORWARD));

#if CUFFT_VERSION >= 10400
// The transform runs on the plan's stream; wait for it before reading back
CUDA_RT_CALL(cudaStreamSynchronize(stream));
#endif

// Scale complex results
float scale{2.f};
scaleComplex(indesc, scale, h_data_out.size(), gpus.size());

// Execute the plan_c2r
CUFFT_CALL(cufftXtExecDescriptor(plan_c2r, indesc, indesc, CUFFT_INVERSE));

#if CUFFT_VERSION >= 10400
// The transform runs on the plan's stream; wait for it before reading back
CUDA_RT_CALL(cudaStreamSynchronize(stream));
#endif

// Copy output data to CPU
CUFFT_CALL(cufftXtMemcpy(plan_c2r, (void *)h_data_out.data(), (void *)indesc,
CUFFT_COPY_DEVICE_TO_HOST));
Expand Down