diff options
| author | Clément Bœsch <clement@stupeflix.com> | 2016-03-25 15:49:09 +0100 |
|---|---|---|
| committer | Clément Bœsch <clement@stupeflix.com> | 2016-03-25 16:14:13 +0100 |
| commit | 277408b7f16eeea60f1f7850638373307f967616 (patch) | |
| tree | fd99716442e177da568943c715d1fb0048854052 /libswscale | |
| parent | 99388eb091d0435a7f4fdb12aaa37f17f2ff951a (diff) | |
sws/aarch64/yuv2rgb: save a few mul and add
27ms to 26ms with UHD 2160 input.
Diffstat (limited to 'libswscale')
| -rw-r--r-- | libswscale/aarch64/yuv2rgb_neon.S | 25 |
1 files changed, 11 insertions, 14 deletions
diff --git a/libswscale/aarch64/yuv2rgb_neon.S b/libswscale/aarch64/yuv2rgb_neon.S index cae5384a97..8cefe2291b 100644 --- a/libswscale/aarch64/yuv2rgb_neon.S +++ b/libswscale/aarch64/yuv2rgb_neon.S @@ -143,20 +143,17 @@ function ff_\ifmt\()_to_\ofmt\()_neon, export=1 load_chroma_\ifmt sub v18.8H, v18.8H, v5.8H // U*(1<<3) - 128*(1<<3) sub v19.8H, v19.8H, v5.8H // V*(1<<3) - 128*(1<<3) - zip1 v6.8H, v19.8H, v19.8H // V1 - zip2 v7.8H, v19.8H, v19.8H // V2 - zip1 v4.8H, v18.8H, v18.8H // U1 - zip2 v5.8H, v18.8H, v18.8H // U2 - sqdmulh v20.8H, v6.8H, v1.H[0] // V1 * v2r (R1) - sqdmulh v21.8H, v7.8H, v1.H[0] // V2 * v2r (R2) - sqdmulh v22.8H, v4.8H, v1.H[1] // U1 * u2g - sqdmulh v23.8H, v5.8H, v1.H[1] // U2 * u2g - sqdmulh v6.8H, v6.8H, v1.H[2] // V1 * v2g - sqdmulh v7.8H, v7.8H, v1.H[2] // V2 * v2g - add v22.8H, v22.8H, v6.8H // U1 * u2g + V1 * v2g (G1) - add v23.8H, v23.8H, v7.8H // U2 * u2g + V2 * v2g (G2) - sqdmulh v24.8H, v4.8H, v1.H[3] // U1 * u2b (B1) - sqdmulh v25.8H, v5.8H, v1.H[3] // U2 * u2b (B2) + sqdmulh v20.8H, v19.8H, v1.H[0] // V * v2r (R) + sqdmulh v22.8H, v18.8H, v1.H[1] // U * u2g + sqdmulh v19.8H, v19.8H, v1.H[2] // V * v2g + add v22.8H, v22.8H, v19.8H // U * u2g + V * v2g (G) + sqdmulh v24.8H, v18.8H, v1.H[3] // U * u2b (B) + zip2 v21.8H, v20.8H, v20.8H // R2 + zip1 v20.8H, v20.8H, v20.8H // R1 + zip2 v23.8H, v22.8H, v22.8H // G2 + zip1 v22.8H, v22.8H, v22.8H // G1 + zip2 v25.8H, v24.8H, v24.8H // B2 + zip1 v24.8H, v24.8H, v24.8H // B1 ld1 {v2.16B}, [x4], #16 // load luma ushll v26.8H, v2.8B, #3 // Y1*(1<<3) ushll2 v27.8H, v2.16B, #3 // Y2*(1<<3) |