Skip to content

Reduce GOP decoder stream synchronizations #33

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
xupinjie merged 1 commit into from
Jun 12, 2026
Merged

Reduce GOP decoder stream synchronizations #33

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
8 changes: 6 additions & 2 deletions ...es/on_demand_video_decoder/ext_impl/src/PyNvOnDemandDecoder/src/PyNvGopDecoder_common.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,8 +109,10 @@ int PyNvGopDecoder::GetYUVFromFrame(NvDecoder* decoder, const uint8_t* pFrame, u
decoded_frame.timestamp = timestamp;
decoded_frame.SetColorRange(color_range);

// Copy the decode frames from device
CUDA_DRVAPI_CALL(cuMemcpyDtoD((CUdeviceptr)pFrame_buffer, (CUdeviceptr)pFrame, decoder->GetFrameSize()));
// Queue the decode-buffer copy on the decoder stream. DecProc synchronizes
// this stream before returning the Python-visible frame.
CUDA_DRVAPI_CALL(cuMemcpyDtoDAsync((CUdeviceptr)pFrame_buffer, (CUdeviceptr)pFrame,
decoder->GetFrameSize(), decoder->GetStream()));

switch (decoded_frame.format) {
case Pixel_Format_NV12: {
Expand Down Expand Up @@ -481,6 +483,8 @@ void PyNvGopDecoder::DecProc(AVColorRange color_range, NvDecoder* decoder,
<< " is different with number of frame id:" << std::to_string(sorted_frame_ids.size());
}

CUDA_DRVAPI_CALL(cuStreamSynchronize(decoder->GetStream()));

nvtxRangePop();
}

Expand Down