[FFmpeg-devel] [PATCH 4/5] lavc/aarch64: Add neon implementation for vsse_intra16
Hubert Mazur
hum at semihalf.com
Mon Aug 22 18:26:26 EEST 2022
Provide optimized implementation for vsse_intra16 for arm64.
Performance tests are shown below.
- vsse_4_c: 153.7
- vsse_4_neon: 34.2
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum at semihalf.com>
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 3 +
libavcodec/aarch64/me_cmp_neon.S | 75 ++++++++++++++++++++++++
2 files changed, 78 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c
index af83f7ed1e..8c295d5457 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -47,6 +47,8 @@ int vsad_intra16_neon(MpegEncContext *c, const uint8_t *s, const uint8_t *dummy,
ptrdiff_t stride, int h) ;
int vsse16_neon(MpegEncContext *c, const uint8_t *s1, const uint8_t *s2,
ptrdiff_t stride, int h);
+int vsse_intra16_neon(MpegEncContext *c, const uint8_t *s, const uint8_t *dummy,
+ ptrdiff_t stride, int h);
av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
{
@@ -69,5 +71,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
c->vsad[4] = vsad_intra16_neon;
c->vsse[0] = vsse16_neon;
+ c->vsse[4] = vsse_intra16_neon;
}
}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index 126e267fdc..46d4dade5d 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -814,3 +814,78 @@ function vsad_intra16_neon, export=1
ret
endfunc
+
+function vsse_intra16_neon, export=1
+ // x0 unused
+ // x1 uint8_t *pix1
+ // x2 uint8_t *dummy
+ // x3 ptrdiff_t stride
+ // w4 int h
+
+ add x5, x1, x3 // pix1 + stride
+ movi v16.4s, #0
+ movi v17.4s, #0
+
+ sub w4, w4, #1 // we need to make h-1 iterations
+ cmp w4, #4
+ b.lt 2f
+
+// make 4 iterations at once
+1:
+ // v = abs( pix1[0] - pix1[0 + stride] )
+ // score = sum( v * v )
+ // abs(x) = ( (x > 0) ? (x) : (-x) )
+
+ ld1 {v0.16b}, [x1], x3
+ ld1 {v1.16b}, [x5], x3
+ ld1 {v2.16b}, [x1], x3
+ uabd v30.16b, v0.16b, v1.16b
+ ld1 {v3.16b}, [x5], x3
+ ld1 {v4.16b}, [x1], x3
+ umull v29.8h, v30.8b, v30.8b
+ umull2 v28.8h, v30.16b, v30.16b
+ uabd v27.16b, v2.16b, v3.16b
+ uadalp v16.4s, v29.8h
+ ld1 {v5.16b}, [x5], x3
+ umull v26.8h, v27.8b, v27.8b
+ uadalp v17.4s, v28.8h
+ ld1 {v6.16b}, [x1], x3
+ umull2 v27.8h, v27.16b, v27.16b
+ uabd v25.16b, v4.16b, v5.16b
+ uadalp v16.4s, v26.8h
+ umull v24.8h, v25.8b, v25.8b
+ ld1 {v7.16b}, [x5], x3
+ uadalp v17.4s, v27.8h
+ umull2 v25.8h, v25.16b, v25.16b
+ uabd v23.16b, v6.16b, v7.16b
+ uadalp v16.4s, v24.8h
+ umull v22.8h, v23.8b, v23.8b
+ uadalp v17.4s, v25.8h
+ umull2 v23.8h, v23.16b, v23.16b
+ sub w4, w4, #4
+ uadalp v16.4s, v22.8h
+ cmp w4, #4
+ uadalp v17.4s, v23.8h
+
+ b.ge 1b
+ cbz w4, 3f
+
+// iterate by one
+2:
+ ld1 {v0.16b}, [x1], x3
+ ld1 {v1.16b}, [x5], x3
+ subs w4, w4, #1
+ uabd v30.16b, v0.16b, v1.16b
+ umull v29.8h, v30.8b, v30.8b
+ umull2 v30.8h, v30.16b, v30.16b
+ uadalp v16.4s, v29.8h
+ uadalp v17.4s, v30.8h
+ cbnz w4, 2b
+
+3:
+ add v16.4s, v16.4s, v17.4S
+ uaddlv d17, v16.4s
+ fmov w0, s17
+
+ ret
+endfunc
--
2.34.1
More information about the ffmpeg-devel
mailing list