coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2026-10-28 18:07:12 +08:00
Code Issues Packages Projects Releases Wiki Activity
libyuv/source
History
Frank Barchard 58ac5cffcb [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 
Optimize 32-bit ARGBRotate (0, 90, 180, 270) across Intel x86/x64
architectures (SSE2, AVX2, AVX512BW), with matching kernels for ARM
(NEON, SVE2, SME), WASM, LoongArch (LSX, LASX), and RISC-V (RVV):

- Replace column-by-column ScaleARGBRowDownEven gather loops in
  ARGBTranspose (90/270) with full-plane horizontal tile transposes:
  - AVX512BW (TransposeWxH_32_AVX512BW / TransposeNx4_32_AVX512BW):
    Transpose 16x16 blocks across zmm0..zmm31 using 4-stage
    vpunpckldq/hdq, vpunpcklqdq/hqdq, and vshufi32x4 butterflies with
    k-masked right and bottom tails, eliminating Any wrappers and scalar
    C edge fallbacks.
  - AVX2 (TransposeWxH_32_AVX2 / TransposeNx4_32_AVX2):
    Transpose 8x8 blocks across ymm0..ymm15 using vpunpck* and
    vperm2i128 with vpmaskmovd masked tails.
  - SSE2 (Transpose4x4_32_SSE2 / Transpose4x4_32_Any_SSE2):
    Transpose 4x4 blocks in xmm registers.
- Implement ARGBRotate180 directly via ARGBMirror with negative height,
  avoiding intermediate row buffer allocations and extra copies.
- Implement ARGBRotate0 via ARGBCopy.

Speedup vs Before (1275x1650, US Letter 150 DPI, 1000 iterations):
CPU                  | 0 deg   | 90 deg  | 180 deg | 270 deg
---------------------|---------|---------|---------|--------
Intel Ice Lake       | 1.02x   | 1.34x   | 1.18x   | 1.33x
Intel Skylake-X      | 1.02x   | 1.20x   | 1.07x   | 1.13x
Intel Emerald Rapids | 1.01x   | 1.07x   | 1.00x   | 1.07x
Intel AVX-512 vs C   | 1.00x   | 2.67x   | 2.41x   | 2.63x

R=rrwinterton@gmail.com

Bug: b/42280915
Test: LIBYUV_WIDTH=1275 LIBYUV_HEIGHT=1650 LIBYUV_REPEAT=1000 ./out/Release/libyuv_unittest --gtest_filter=*ARGBRotate*
Change-Id: Ib308a225c17e46eab083a9da4033598fcf0c303c
2026-10-08 03:05:41 -07:00
..
compare_common.cc [libyuv] Keep LIBYUV_API in basic_types.h and remove legacy types and bool macros 2026-08-28 08:32:19 -07:00
compare_gcc.cc Compare AVX512BW and AVX2, intrinsics for Visual C and ROWWIN 2026-10-02 15:30:05 -07:00
compare_neon64.cc [libyuv] Keep LIBYUV_API in basic_types.h and remove legacy types and bool macros 2026-08-28 08:32:19 -07:00
compare_neon.cc [libyuv] Keep LIBYUV_API in basic_types.h and remove legacy types and bool macros 2026-08-28 08:32:19 -07:00
compare_wasm.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
compare_win.cc Compare AVX512BW and AVX2, intrinsics for Visual C and ROWWIN 2026-10-02 15:30:05 -07:00
compare.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
convert_argb.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
convert_from_argb.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
convert_from.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
convert_jpeg.cc [libyuv] Keep LIBYUV_API in basic_types.h and remove legacy types and bool macros 2026-08-28 08:32:19 -07:00
convert_to_argb.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
convert_to_i420.cc clang format and bump version 2026-09-28 10:42:45 -07:00
convert.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
cpu_id.cc [libyuv] Add WASM SIMD128 and Relaxed SIMD for RAWToRGB24Row 2026-10-05 17:45:41 -07:00
mjpeg_decoder.cc [libyuv] Fix RISC-V CMake/GN/Bazel builds and C compiler driver linking 2026-09-28 10:59:30 -07:00
mjpeg_validate.cc [libyuv] Keep LIBYUV_API in basic_types.h and remove legacy types and bool macros 2026-08-28 08:32:19 -07:00
planar_functions.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_any.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_argb.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_common.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_gcc.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_lasx.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_lsx.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_neon64.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_neon.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_rvv.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_sme.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_sve.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_wasm.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate_win.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
rotate.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
row_any.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
row_common.cc clang format and bump version 2026-09-28 10:42:45 -07:00
row_gcc.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
row_lasx.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
row_lsx.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
row_neon64.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
row_neon.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
row_rvv.cc [libyuv] Optimize ARGBRotate for Intel SSE2, AVX2, and AVX-512 2026-10-08 03:05:41 -07:00
row_sme.cc Convert8To16 takes bits instead of scale. 2026-09-01 16:07:23 -07:00
row_sve.cc clang format and bump version 2026-09-28 10:42:45 -07:00
row_wasm.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
row_win.cc [libyuv] AVX512, NEON, WASM and RVV alpha copy and mirror rows 2026-10-07 12:44:53 -07:00
scale_any.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
scale_argb.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
scale_common.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
scale_gcc.cc [libyuv] Add ScaleRowDown2Box_AVX512BW with tail masking 2026-10-05 10:56:34 -07:00
scale_lsx.cc DetilePlane and unittest for NEON 2022-01-31 20:05:55 +00:00
scale_neon64.cc Apply format with no code changes 2025-02-24 23:57:01 -08:00
scale_neon.cc Apply format with no code changes 2025-02-24 23:57:01 -08:00
scale_rgb.cc Fix Coverity Control flow issues (DEADCODE) 2026-07-01 10:39:22 -07:00
scale_rvv.cc [libyuv] Clobber "vl" in RVV inline assembly only for GCC 2026-10-06 12:19:58 -07:00
scale_sme.cc Apply clang format 2025-01-02 13:31:20 -08:00
scale_uv.cc Remove the Abs() function in source/scale_uv.cc 2026-10-07 16:10:15 -07:00
scale_wasm.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
scale_win.cc BGRAToI420 use BgraConstants for a direct conversion using AVX512BW 2026-06-08 12:21:47 -07:00
scale.cc [libyuv] Add WASM SIMD128 compare, rotate, scale and row functions 2026-10-06 19:05:05 -07:00
test.sh Optimze ABGRToI420 for AVX2 2020-06-04 18:24:45 +00:00
video_common.cc Lint cleanup after C99 change CL 2018-01-24 19:16:03 +00:00