Parallelize CustomWaveform and stbi_write_png using OpenMP #8
Conversation
Implemented OpenMP parallelization for: 1. CustomWaveform::Draw: Added a parallel path for waveforms without per-point code, replicating the default vertex generation logic to avoid shared state contention. 2. CustomWaveform::SmoothWave: Parallelized the smoothing loop by calculating dependency indices locally. 3. stbi_write_png (SOIL2): Parallelized the PNG line filtering loop, ensuring thread-local scratch buffers. Also updated vendor/SOIL2/CMakeLists.txt to link OpenMP::OpenMP_C when enabled.
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
ford442
deleted the
feature/openmp-custom-wave-stbi-3007410145637072838
branch
1 participant
This PR implements OpenMP parallelization for
CustomWaveformrendering andstb_image_writePNG encoding, as suggested indocs/openmp.md.Changes:
vendor/SOIL2/src/SOIL2/stb_image_write.h: Parallelized the main row loop instbi_write_png_to_memusing#pragma omp parallel for. Movedline_bufferallocation inside the parallel region to ensure thread safety. Added failure handling via atomic flag.vendor/SOIL2/CMakeLists.txt: Added logic to linkOpenMP::OpenMP_Cand definePRJM_ENABLE_OPENMPfor theSOIL2target ifENABLE_OPENMPis ON.src/libprojectM/MilkdropPreset/CustomWaveform.cpp:Draw, added a check for!m_perPointContext.perPointCodeHandle. If no per-point code is present, a parallel loop generates vertices using local logic (bypassing the sharedPerPointContext).SmoothWave, parallelized the loop by removing stateful dependency variables (iBelow,iAbove2) and calculating them directly from the loop index.These changes are guarded by
PRJM_ENABLE_OPENMP.PR created automatically by Jules for task 3007410145637072838 started by @ford442