diff options
| author | Grzegorz Bernacki <gjb@semihalf.com> | 2022-10-03 16:10:19 +0200 |
|---|---|---|
| committer | Martin Storsjö <martin@martin.st> | 2022-10-04 13:24:20 +0300 |
| commit | 8f4b000c37353807666db3f5025cc4a77f039c32 (patch) | |
| tree | 04569fefe9b6e6e8063e442334d40e4a241b65cf /libavcodec/aarch64/me_cmp_neon.S | |
| parent | bad67cb9fd53885ee2e5fba4db28b1f1bb315a8f (diff) | |
| download | ffmpeg-8f4b000c37353807666db3f5025cc4a77f039c32.tar.gz | |
lavc/aarch64: Add neon implementation for vsse_intra8
Provide optimized implementation for vsse_intra8 for arm64.
Performance tests are shown below.
- vsse_5_c: 87.7
- vsse_5_neon: 26.2
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Co-authored-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Martin Storsjö <martin@martin.st>
Diffstat (limited to 'libavcodec/aarch64/me_cmp_neon.S')
| -rw-r--r-- | libavcodec/aarch64/me_cmp_neon.S | 53 |
1 files changed, 53 insertions, 0 deletions
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S index 104e02f495..d8a18cd4b8 100644 --- a/libavcodec/aarch64/me_cmp_neon.S +++ b/libavcodec/aarch64/me_cmp_neon.S @@ -1106,6 +1106,59 @@ function vsse_intra16_neon, export=1 ret endfunc +function vsse_intra8_neon, export=1 + // x0 unused + // x1 uint8_t *pix1 + // x2 uint8_t *dummy + // x3 ptrdiff_t stride + // w4 int h + + sub w4, w4, #1 // we need to make h-1 iterations + ld1 {v0.8b}, [x1], x3 + cmp w4, #3 + movi v16.4s, #0 + + b.lt 2f + +1: + // v = abs( pix1[0] - pix1[0 + stride] ) + // score = sum( v * v ) + ld1 {v1.8b}, [x1], x3 + ld1 {v2.8b}, [x1], x3 + uabd v30.8b, v0.8b, v1.8b + ld1 {v3.8b}, [x1], x3 + uabd v27.8b, v1.8b, v2.8b + umull v29.8h, v30.8b, v30.8b + uabd v25.8b, v2.8b, v3.8b + umull v26.8h, v27.8b, v27.8b + uadalp v16.4s, v29.8h + umull v24.8h, v25.8b, v25.8b + uadalp v16.4s, v26.8h + sub w4, w4, #3 + uadalp v16.4s, v24.8h + cmp w4, #3 + mov v0.8b, v3.8b + + b.ge 1b + cbz w4, 3f + +// iterate by one +2: + ld1 {v1.8b}, [x1], x3 + subs w4, w4, #1 + uabd v30.8b, v0.8b, v1.8b + mov v0.8b, v1.8b + umull v29.8h, v30.8b, v30.8b + uadalp v16.4s, v29.8h + cbnz w4, 2b + +3: + uaddlv d17, v16.4s + fmov w0, s17 + + ret +endfunc + function nsse16_neon, export=1 // x0 multiplier // x1 uint8_t *pix1 |