Add AVX2 intra prediction for 16x16 luma and 8x8 chroma#3934
Open
rbenaley wants to merge 1 commit into
Open
Conversation
Implement AVX2-optimized intra prediction for the four 16x16 luma modes (V, H, DC, Plane) and 8x8 chroma vertical prediction. V/DC use vbroadcasti128/vpbroadcastb with 256-bit stores (8 instead of 16). H uses vpbroadcastb for single-instruction byte broadcast. Plane processes all 16 pixels per row in one pass via vpmullw+vpaddw+vpackuswb+vpermq. Chroma V uses vpbroadcastq to fill the 64-byte block with 2 stores. All code is guarded by %ifdef HAVE_AVX2 / WELS_CPU_AVX2 and selected at runtime via CPUID detection.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This adds AVX2-optimized intra prediction for the four 16x16 luma modes (Vertical, Horizontal, DC, Plane) and 8x8 chroma vertical prediction.
Key improvements over the SSE2 versions:
vbroadcasti128/vpbroadcastb+ 256-bit stores write two rows per operation (8 stores instead of 16)vpbroadcastb xmm, [mem]replaces a multi-instruction broadcast sequencevpmullw+vpaddw+vpackuswb+vpermq), versus two 8-pixel passes in SSE2vpbroadcastqfills the entire 64-byte block with just 2 storesFiles changed:
codec/common/x86/intra_pred_com.asm— V, H functions (+44 lines)codec/encoder/core/x86/intra_pred.asm— DC, Plane, Chroma V (+215 lines)codec/common/inc/intra_pred_common.h— declarations (+4 lines)codec/encoder/core/inc/get_intra_predictor.h— declarations (+6 lines)codec/encoder/core/src/get_intra_predictor.cpp— registration (+9 lines)These optimizations were developed for Vauban, an open-source privileged access management (PAM) bastion that uses OpenH264 for real-time H.264 encoding of RDP desktop sessions streamed to web browsers. Combined with AVX2 SAD functions, enabling full AVX2 support reduced CPU usage per session by approximately 50% on an Intel Xeon E-2246G running FreeBSD.
For a detailed technical writeup covering the encoding pipeline context, implementation choices, and performance measurements, see:
https://github.com/rbenaley/Vauban/blob/main/docs/technical/Vauban_OpenH264_AVX2_Optimizations_EN(1.0).md