[FFmpeg-cvslog] lavc/aarch64: Add neon implementation for vsad_intra16
ffmpeg | branch: master | Hubert Mazur | Thu Sep 8
11:25:05 2022 +0200| [ce03ea3e796bdf9013da51d53e89759a92707c4a] | committer:
Martin Storsjö
lavc/aarch64: Add neon implementation for vsad_intra16
Provide optimized implementation for vsad_intra16 function for arm64.
Performance comparison tests are shown below.
- vsad_4_c: 177.5
- vsad_4_neon: 23.5
Benchmarks and tests are run with checkasm tool on AWS Gravtion 3.
Signed-off-by: Hubert Mazur
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=ce03ea3e796bdf9013da51d53e89759a92707c4a
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++
libavcodec/aarch64/me_cmp_neon.S | 48
2 files changed, 51 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c
b/libavcodec/aarch64/me_cmp_init_aarch64.c
index 7b81e48d16..af83f7ed1e 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -43,6 +43,8 @@ int sse4_neon(MpegEncContext *v, const uint8_t *pix1, const
uint8_t *pix2,
int vsad16_neon(MpegEncContext *c, const uint8_t *s1, const uint8_t *s2,
ptrdiff_t stride, int h);
+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);
@@ -64,6 +66,7 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c,
AVCodecContext *avctx)
c->sse[2] = sse4_neon;
c->vsad[0] = vsad16_neon;
+c->vsad[4] = vsad_intra16_neon;
c->vsse[0] = vsse16_neon;
}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index b3f376aa60..ce198ea227 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -736,3 +736,51 @@ function vsse16_neon, export=1
ret
endfunc
+
+function vsad_intra16_neon, export=1
+// x0 unused
+// x1 uint8_t *pix1
+// x2 uint8_t *dummy
+// x3 ptrdiff_t stride
+// w4 int h
+
+ld1 {v0.16b}, [x1], x3
+sub w4, w4, #1 // we need to make h-1 iterations
+cmp w4, #3
+moviv16.8h, #0
+b.lt2f
+
+// make 4 iterations at once
+1:
+// v = abs( pix1[0] - pix1[0 + stride] )
+// score = sum(v)
+ld1 {v1.16b}, [x1], x3
+ld1 {v2.16b}, [x1], x3
+uabal v16.8h, v0.8b, v1.8b
+ld1 {v3.16b}, [x1], x3
+uabal2 v16.8h, v0.16b, v1.16b
+sub w4, w4, #3
+uabal v16.8h, v1.8b, v2.8b
+cmp w4, #3
+uabal2 v16.8h, v1.16b, v2.16b
+mov v0.16b, v3.16b
+uabal v16.8h, v2.8b, v3.8b
+uabal2 v16.8h, v2.16b, v3.16b
+b.ge1b
+cbz w4, 3f
+
+// iterate by one
+2:
+ld1 {v1.16b}, [x1], x3
+subsw4, w4, #1
+uabal v16.8h, v0.8b, v1.8b
+uabal2 v16.8h, v0.16b, v1.16b
+mov v0.16b, v1.16b
+cbnzw4, 2b
+
+3:
+uaddlv s17, v16.8h
+fmovw0, s17
+
+ret
+endfunc
___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
To unsubscribe, visit link above, or email
ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] lavc/aarch64: Add neon implementation for vsad16
ffmpeg | branch: master | Hubert Mazur | Thu Sep 8
11:25:03 2022 +0200| [200f5e578f2fbf70a966f08257e0500a6f1ddd6c] | committer:
Martin Storsjö
lavc/aarch64: Add neon implementation for vsad16
Provide optimized implementation of vsad16 function for arm64.
Performance comparison tests are shown below.
- vsad_0_c: 285.2
- vsad_0_neon: 39.5
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Co-authored-by: Martin Storsjö
Signed-off-by: Hubert Mazur
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=200f5e578f2fbf70a966f08257e0500a6f1ddd6c
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 5 +++
libavcodec/aarch64/me_cmp_neon.S | 65
2 files changed, 70 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c
b/libavcodec/aarch64/me_cmp_init_aarch64.c
index fb7c3f5059..ddc5d05611 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -41,6 +41,9 @@ int sse8_neon(MpegEncContext *v, const uint8_t *pix1, const
uint8_t *pix2,
int sse4_neon(MpegEncContext *v, const uint8_t *pix1, const uint8_t *pix2,
ptrdiff_t stride, int h);
+int vsad16_neon(MpegEncContext *c, const uint8_t *s1, const uint8_t *s2,
+ptrdiff_t stride, int h);
+
av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
{
int cpu_flags = av_get_cpu_flags();
@@ -57,5 +60,7 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c,
AVCodecContext *avctx)
c->sse[0] = sse16_neon;
c->sse[1] = sse8_neon;
c->sse[2] = sse4_neon;
+
+c->vsad[0] = vsad16_neon;
}
}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index 4198985c6c..1d0b166d69 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -584,3 +584,68 @@ function sse4_neon, export=1
ret
endfunc
+
+function vsad16_neon, export=1
+// x0 unused
+// x1 uint8_t *pix1
+// x2 uint8_t *pix2
+// x3 ptrdiff_t stride
+// w4 int h
+
+ld1 {v0.16b}, [x1], x3 // Load pix1[0], first
iteration
+ld1 {v1.16b}, [x2], x3 // Load pix2[0], first
iteration
+
+sub w4, w4, #1 // we need to make h-1
iterations
+moviv16.8h, #0
+
+cmp w4, #3 // check if we can
make 3 iterations at once
+usubl v31.8h, v0.8b, v1.8b// Signed difference
pix1[0] - pix2[0], first iteration
+usubl2 v30.8h, v0.16b, v1.16b // Signed difference
pix1[0] - pix2[0], first iteration
+
+b.lt2f
+
+1:
+// abs(pix1[0] - pix2[0] - pix1[0 + stride] + pix2[0 + stride])
+ld1 {v0.16b}, [x1], x3 // Load pix1[0 +
stride], first iteration
+ld1 {v1.16b}, [x2], x3 // Load pix2[0 +
stride], first iteration
+ld1 {v2.16b}, [x1], x3 // Load pix1[0 +
stride], second iteration
+ld1 {v3.16b}, [x2], x3 // Load pix2[0 +
stride], second iteration
+usubl v29.8h, v0.8b, v1.8b // Signed difference
pix1[0 + stride] - pix2[0 + stride], first iteration
+usubl2 v28.8h, v0.16b, v1.16b // Signed difference
pix1[0 + stride] - pix2[0 + stride], first iteration
+ld1 {v4.16b}, [x1], x3 // Load pix1[0 +
stride], third iteration
+ld1 {v5.16b}, [x2], x3 // Load pix2[0 +
stride], third iteration
+usubl v27.8h, v2.8b, v3.8b // Signed difference
pix1[0 + stride] - pix2[0 + stride], second iteration
+sabav16.8h, v31.8h, v29.8h // Signed absolute
difference and accumulate the result. first iteration
+usubl2 v26.8h, v2.16b, v3.16b // Signed difference
pix1[0 + stride] - pix2[0 + stride], second iteration
+sabav16.8h, v30.8h, v28.8h // Signed absolute
difference and accumulate the result. first iteration
+usubl v25.8h, v4.8b, v5.8b // Signed difference
pix1[0 + stride] - pix2[0 + stride], third iteration
+usubl2 v24.8h, v4.16b, v5.16b // Signed difference
pix1[0 + stride] - pix2[0 + stride], third iteration
+sabav16.8h, v29.8h, v27.8h // Signed absolute
difference and accumulate the result. second iteration
+mov v31.16b, v25.16b
+sabav16.8h, v28.8h, v26.8h // Signed absolute
difference and accumulate the result. second iteration
+sub w4, w4, #3 // h -= 3
+mov
[FFmpeg-cvslog] lavc/aarch64: Add neon implementation of vsse16
ffmpeg | branch: master | Hubert Mazur | Thu Sep 8
11:25:04 2022 +0200| [c495a4b32d352e087318d3a09a9bb4f2b55cfa04] | committer:
Martin Storsjö
lavc/aarch64: Add neon implementation of vsse16
Provide optimized implementation of vsse16 for arm64.
Performance comparison tests are shown below.
- vsse_0_c: 257.7
- vsse_0_neon: 59.2
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c495a4b32d352e087318d3a09a9bb4f2b55cfa04
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 4 ++
libavcodec/aarch64/me_cmp_neon.S | 87
2 files changed, 91 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c
b/libavcodec/aarch64/me_cmp_init_aarch64.c
index ddc5d05611..7b81e48d16 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -43,6 +43,8 @@ int sse4_neon(MpegEncContext *v, const uint8_t *pix1, const
uint8_t *pix2,
int vsad16_neon(MpegEncContext *c, const uint8_t *s1, const uint8_t *s2,
ptrdiff_t stride, int h);
+int vsse16_neon(MpegEncContext *c, const uint8_t *s1, const uint8_t *s2,
+ptrdiff_t stride, int h);
av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
{
@@ -62,5 +64,7 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c,
AVCodecContext *avctx)
c->sse[2] = sse4_neon;
c->vsad[0] = vsad16_neon;
+
+c->vsse[0] = vsse16_neon;
}
}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index 1d0b166d69..b3f376aa60 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -649,3 +649,90 @@ function vsad16_neon, export=1
ret
endfunc
+
+function vsse16_neon, export=1
+// x0 unused
+// x1 uint8_t *pix1
+// x2 uint8_t *pix2
+// x3 ptrdiff_t stride
+// w4 int h
+
+ld1 {v0.16b}, [x1], x3 // Load pix1[0], first
iteration
+ld1 {v1.16b}, [x2], x3 // Load pix2[0], first
iteration
+
+sub w4, w4, #1 // we need to make h-1
iterations
+moviv16.4s, #0
+moviv17.4s, #0
+
+cmp w4, #3 // check if we can
make 3 iterations at once
+usubl v31.8h, v0.8b, v1.8b// Signed difference
of pix1[0] - pix2[0], first iteration
+usubl2 v30.8h, v0.16b, v1.16b // Signed difference
of pix1[0] - pix2[0], first iteration
+b.le2f
+
+
+1:
+// x = abs(pix1[0] - pix2[0] - pix1[0 + stride] + pix2[0 + stride])
+// res = (x) * (x)
+ld1 {v0.16b}, [x1], x3 // Load pix1[0 +
stride], first iteration
+ld1 {v1.16b}, [x2], x3 // Load pix2[0 +
stride], first iteration
+ld1 {v2.16b}, [x1], x3 // Load pix1[0 +
stride], second iteration
+ld1 {v3.16b}, [x2], x3 // Load pix2[0 +
stride], second iteration
+usubl v29.8h, v0.8b, v1.8b
+usubl2 v28.8h, v0.16b, v1.16b
+ld1 {v4.16b}, [x1], x3 // Load pix1[0 +
stride], third iteration
+ld1 {v5.16b}, [x2], x3 // Load pix1[0 +
stride], third iteration
+sabdv31.8h, v31.8h, v29.8h
+sabdv30.8h, v30.8h, v28.8h
+usubl v27.8h, v2.8b, v3.8b
+usubl2 v26.8h, v2.16b, v3.16b
+usubl v25.8h, v4.8b, v5.8b
+usubl2 v24.8h, v4.16b, v5.16b
+sabdv29.8h, v29.8h, v27.8h
+sabdv27.8h, v27.8h, v25.8h
+umlal v16.4s, v31.4h, v31.4h
+umlal2 v17.4s, v31.8h, v31.8h
+sabdv28.8h, v28.8h, v26.8h
+sabdv26.8h, v26.8h, v24.8h
+umlal v16.4s, v30.4h, v30.4h
+umlal2 v17.4s, v30.8h, v30.8h
+mov v31.16b, v25.16b
+umlal v16.4s, v29.4h, v29.4h
+umlal2 v17.4s, v29.8h, v29.8h
+mov v30.16b, v24.16b
+umlal v16.4s, v28.4h, v28.4h
+umlal2 v17.4s, v28.8h, v28.8h
+sub w4, w4, #3
+umlal v16.4s, v27.4h, v27.4h
+umlal2 v17.4s, v27.8h, v27.8h
+cmp w4, #3
+umlal v16.4s, v26.4h, v26.4h
+umlal2 v17.4s, v26.8h, v26.8h
+
+b.ge1b
+
+cbz w4, 3f
+
+// iterate by once
+2:
+ld1 {v0.16b}, [x1], x3
+ld1 {v1.16b}, [x2], x3
+subs
[FFmpeg-cvslog] lavc/aarch64: Provide neon implementation of nsse16
ffmpeg | branch: master | Hubert Mazur | Thu Sep 8
11:25:07 2022 +0200| [06b98e396adc467a5164a03d71dd71508a2d8881] | committer:
Martin Storsjö
lavc/aarch64: Provide neon implementation of nsse16
Add vectorized implementation of nsse16 function.
Performance comparison tests are shown below.
- nsse_0_c: 682.2
- nsse_0_neon: 116.5
Benchmarks and tests run with checkasm tool on AWS Graviton 3.
Co-authored-by: Martin Storsjö
Signed-off-by: Hubert Mazur
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=06b98e396adc467a5164a03d71dd71508a2d8881
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 15
libavcodec/aarch64/me_cmp_neon.S | 122 +++
2 files changed, 137 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c
b/libavcodec/aarch64/me_cmp_init_aarch64.c
index 8c295d5457..ade3e9a4c1 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -49,6 +49,10 @@ 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);
+int nsse16_neon(int multiplier, const uint8_t *s, const uint8_t *s2,
+ptrdiff_t stride, int h);
+int nsse16_neon_wrapper(MpegEncContext *c, const uint8_t *s1, const uint8_t
*s2,
+ptrdiff_t stride, int h);
av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
{
@@ -72,5 +76,16 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c,
AVCodecContext *avctx)
c->vsse[0] = vsse16_neon;
c->vsse[4] = vsse_intra16_neon;
+
+c->nsse[0] = nsse16_neon_wrapper;
}
}
+
+int nsse16_neon_wrapper(MpegEncContext *c, const uint8_t *s1, const uint8_t
*s2,
+ptrdiff_t stride, int h)
+{
+if (c)
+return nsse16_neon(c->avctx->nsse_weight, s1, s2, stride, h);
+else
+return nsse16_neon(8, s1, s2, stride, h);
+}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index cf2b8da425..f8998749a5 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -847,3 +847,125 @@ function vsse_intra16_neon, export=1
ret
endfunc
+
+function nsse16_neon, export=1
+// x0 multiplier
+// x1 uint8_t *pix1
+// x2 uint8_t *pix2
+// x3 ptrdiff_t stride
+// w4 int h
+
+str x0, [sp, #-0x40]!
+stp x1, x2, [sp, #0x10]
+stp x3, x4, [sp, #0x20]
+str x30, [sp, #0x30]
+bl X(sse16_neon)
+ldr x30, [sp, #0x30]
+mov w9, w0 // here we
store score1
+ldr x5, [sp]
+ldp x1, x2, [sp, #0x10]
+ldp x3, x4, [sp, #0x20]
+add sp, sp, #0x40
+
+moviv16.8h, #0
+moviv17.8h, #0
+moviv18.8h, #0
+moviv19.8h, #0
+
+ld1 {v0.16b}, [x1], x3
+subsw4, w4, #1 // we need to
make h-1 iterations
+ld1 {v2.16b}, [x2], x3
+ext v1.16b, v0.16b, v0.16b, #1 // x1 + 1
+cmp w4, #2
+ext v3.16b, v2.16b, v2.16b, #1 // x2 + 1
+
+b.lt2f
+
+// make 2 iterations at once
+1:
+ld1 {v4.16b}, [x1], x3
+ld1 {v6.16b}, [x2], x3
+ld1 {v20.16b}, [x1], x3
+ext v5.16b, v4.16b, v4.16b, #1 // x1 + stride
+ 1
+usubl v31.8h, v0.8b, v4.8b
+usubl2 v30.8h, v0.16b, v4.16b
+ld1 {v22.16b}, [x2], x3
+usubl v29.8h, v1.8b, v5.8b
+usubl2 v28.8h, v1.16b, v5.16b
+ext v7.16b, v6.16b, v6.16b, #1 // x2 + stride
+ 1
+sabav16.8h, v31.8h, v29.8h
+ext v21.16b, v20.16b, v20.16b, #1
+sabav17.8h, v30.8h, v28.8h
+usubl v27.8h, v2.8b, v6.8b
+usubl2 v26.8h, v2.16b, v6.16b
+ext v23.16b, v22.16b, v22.16b, #1
+usubl v25.8h, v3.8b, v7.8b
+usubl2 v24.8h, v3.16b, v7.16b
+sabav18.8h, v27.8h, v25.8h
+sabav19.8h, v26.8h, v24.8h
+
+usubl v31.8h, v4.8b, v20.8b
+usubl2 v30.8h, v4.16b, v20.16b
+usubl v29.8h, v5.8b, v21.8b
+usubl2 v28.8h, v5.16b, v21.16b
+sabav16.8h, v31.8h, v29.8h
+sabav1
[FFmpeg-cvslog] lavc/aarch64: Add neon implementation for vsse_intra16
ffmpeg | branch: master | Hubert Mazur | Thu Sep 8
11:25:06 2022 +0200| [908abe8032d2e56f9b94a7ae387e415de4c29115] | committer:
Martin Storsjö
lavc/aarch64: Add neon implementation for vsse_intra16
Provide optimized implementation for vsse_intra16 for arm64.
Performance tests are shown below.
- vsse_4_c: 155.2
- vsse_4_neon: 36.2
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=908abe8032d2e56f9b94a7ae387e415de4c29115
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++
libavcodec/aarch64/me_cmp_neon.S | 63
2 files changed, 66 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 ce198ea227..cf2b8da425 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -784,3 +784,66 @@ 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
+
+ld1 {v0.16b}, [x1], x3
+moviv16.4s, #0
+moviv17.4s, #0
+
+sub w4, w4, #1 // we need to make h-1 iterations
+cmp w4, #3
+b.lt2f
+
+1:
+// v = abs( pix1[0] - pix1[0 + stride] )
+// score = sum( v * v )
+ld1 {v1.16b}, [x1], x3
+ld1 {v2.16b}, [x1], x3
+uabdv30.16b, v0.16b, v1.16b
+ld1 {v3.16b}, [x1], x3
+umull v29.8h, v30.8b, v30.8b
+umull2 v28.8h, v30.16b, v30.16b
+uabdv27.16b, v1.16b, v2.16b
+uadalp v16.4s, v29.8h
+umull v26.8h, v27.8b, v27.8b
+umull2 v27.8h, v27.16b, v27.16b
+uadalp v17.4s, v28.8h
+uabdv25.16b, v2.16b, v3.16b
+uadalp v16.4s, v26.8h
+umull v24.8h, v25.8b, v25.8b
+umull2 v25.8h, v25.16b, v25.16b
+uadalp v17.4s, v27.8h
+sub w4, w4, #3
+uadalp v16.4s, v24.8h
+cmp w4, #3
+uadalp v17.4s, v25.8h
+mov v0.16b, v3.16b
+
+b.ge1b
+cbz w4, 3f
+
+// iterate by one
+2:
+ld1 {v1.16b}, [x1], x3
+subsw4, w4, #1
+uabdv30.16b, v0.16b, v1.16b
+mov 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
+cbnzw4, 2b
+
+3:
+add v16.4s, v16.4s, v17.4S
+uaddlv d17, v16.4s
+fmovw0, s17
+
+ret
+endfunc
___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
To unsubscribe, visit link above, or email
ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] tests/fate-run: Allow to set input options for encoding pass
ffmpeg | branch: master | Andreas Rheinhardt |
Wed Sep 7 22:36:45 2022 +0200| [a5ab4be081aee22e675d0e78aa9ca0d08f4a5d6f] |
committer: Andreas Rheinhardt
tests/fate-run: Allow to set input options for encoding pass
This will be useful in the next commit.
Signed-off-by: Andreas Rheinhardt
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a5ab4be081aee22e675d0e78aa9ca0d08f4a5d6f
---
tests/fate-run.sh | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/tests/fate-run.sh b/tests/fate-run.sh
index 4008bcbc16..61cc59acc0 100755
--- a/tests/fate-run.sh
+++ b/tests/fate-run.sh
@@ -247,12 +247,13 @@ transcode(){
ffprobe_opts=$6
additional_input=$7
final_decode=$8
+enc_opt_in=$9
test -z "$additional_input" || additional_input="$DEC_OPTS
$additional_input"
encfile="${outdir}/${test}.${enc_fmt}"
test $keep -ge 1 || cleanfiles="$cleanfiles $encfile"
tsrcfile=$(target_path $srcfile)
tencfile=$(target_path $encfile)
-ffmpeg -f $src_fmt $DEC_OPTS -i $tsrcfile $additional_input \
+ffmpeg -f $src_fmt $DEC_OPTS $enc_opt_in -i $tsrcfile $additional_input \
$ENC_OPTS $enc_opt $FLAGS -f $enc_fmt -y $tencfile || return
do_md5sum $encfile
echo $(wc -c $encfile)
___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
To unsubscribe, visit link above, or email
ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] fate/matroska: Add test for updating AV1 extradata
ffmpeg | branch: master | Andreas Rheinhardt | Wed Sep 7 23:57:15 2022 +0200| [91e9a6df33d8b14577fe1ec9623d9d0466fdd7d3] | committer: Andreas Rheinhardt fate/matroska: Add test for updating AV1 extradata Signed-off-by: Andreas Rheinhardt > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=91e9a6df33d8b14577fe1ec9623d9d0466fdd7d3 --- tests/fate/matroska.mak | 6 ++ tests/ref/fate/webm-av1-extradata-update | 32 2 files changed, 38 insertions(+) diff --git a/tests/fate/matroska.mak b/tests/fate/matroska.mak index 63e81f121b..39137ad4be 100644 --- a/tests/fate/matroska.mak +++ b/tests/fate/matroska.mak @@ -49,6 +49,12 @@ FATE_MATROSKA-$(call ALLYES, FLAC_DECODER FLAC_ENCODER FLAC_PARSER \ fate-matroska-flac-extradata-update: CMD = transcode matroska $(TARGET_SAMPLES)/mkv/flac_channel_layouts.mka \ matroska "-map 0 -map 0:0 -c flac -frames:a:2 8" "-map 0 -c copy" +# This tests that the Matroska/WebM muxer writes the AV1 CodecPrivate +# via extradata obtained from packet side data. It also tests that +# the aspect ratio is only written with pixels as DisplayUnit for WebM. +FATE_MATROSKA-$(call REMUX, WEBM MATROSKA, IVF_DEMUXER AV1_PARSER EXTRACT_EXTRADATA_BSF) += fate-webm-av1-extradata-update +fate-webm-av1-extradata-update: CMD = transcode ivf $(TARGET_SAMPLES)/av1/decode_model.ivf webm "-c copy -bsf extract_extradata -sar 3:1" "-c copy" "" "" "-nofind_stream_info" "-nofind_stream_info" + # This test tests demuxing Vorbis and chapters from ogg and muxing it in and # demuxing it from Matroska/WebM. It furthermore tests the WebM muxer, in # particular its DASH mode. Finally, it tests writing the Cues at the front. diff --git a/tests/ref/fate/webm-av1-extradata-update b/tests/ref/fate/webm-av1-extradata-update new file mode 100644 index 00..9dd2056e0e --- /dev/null +++ b/tests/ref/fate/webm-av1-extradata-update @@ -0,0 +1,32 @@ +fbf3091fdf05b2856c578e7c948d68c3 *tests/data/fate/webm-av1-extradata-update.webm +23048 tests/data/fate/webm-av1-extradata-update.webm +#extradata 0: 35, 0x527207cd +#tb 0: 1/1000 +#media_type 0: video +#codec_id 0: av1 +#dimensions 0: 240x100 +#sar 0: 3/1 +0, 0, 0,0, 8168, 0x1851ab62 +0, 42, 42,0, 7040, 0x967788f9, F=0x0 +0, 83, 83,0,4, 0x01f400e2, F=0x0 +0,125,125,0, 48, 0x49ad107e, F=0x0 +0,167,167,0,4, 0x021c00fa, F=0x0 +0,208,208,0, 279, 0x69728439, F=0x0 +0,250,250,0,4, 0x01c400d2, F=0x0 +0,292,292,0, 63, 0x9bbf1836, F=0x0 +0,333,333,0,4, 0x026c012a, F=0x0 +0,375,375,0, 1065, 0xce2003ac, F=0x0 +0,417,417,0,4, 0x019400c2, F=0x0 +0,458,458,0, 52, 0x7a0112f1, F=0x0 +0,500,500,0,4, 0x021c010a, F=0x0 +0,542,542,0, 689, 0x1e8b49e7, F=0x0 +0,583,583,0,4, 0x01e400f2, F=0x0 +0,625,625,0, 209, 0x124c6790, F=0x0 +0,667,667,0, 42, 0xea690e31, F=0x0 +0,708,708,0, 3521, 0xd76ee284, F=0x0 +0,750,750,0, 63, 0x4572188f, F=0x0 +0,792,792,0, 386, 0xb078c259, F=0x0 +0,833,833,0, 178, 0x1ebb5121, F=0x0 +0,875,875,0, 60, 0x729317f7, F=0x0 +0,917,917,0, 40, 0xad970a66, F=0x0 +0,958,958,0, 61, 0xcc0d1a20, F=0x0 ___ ffmpeg-cvslog mailing list ffmpeg-cvslog@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog To unsubscribe, visit link above, or email ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] lavu/tx: add the inplace flag to PFA FFTs
ffmpeg | branch: master | Lynne | Sat Sep 10 02:26:49 2022
+0200| [645a1f4422ad9c8c954e7c42bef2281cac96ab18] | committer: Lynne
lavu/tx: add the inplace flag to PFA FFTs
They support in-place, because they have to use a temporary buffer.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=645a1f4422ad9c8c954e7c42bef2281cac96ab18
---
libavutil/tx_template.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavutil/tx_template.c b/libavutil/tx_template.c
index 35b61fa477..542c15e480 100644
--- a/libavutil/tx_template.c
+++ b/libavutil/tx_template.c
@@ -836,7 +836,7 @@ static const FFTXCodelet TX_NAME(ff_tx_fft_pfa_##N##xM_def)
= {\
.name = TX_NAME_STR("fft_pfa_" #N "xM"),
\
.function = TX_NAME(ff_tx_fft_pfa_##N##xM),
\
.type = TX_TYPE(FFT),
\
-.flags = AV_TX_UNALIGNED | FF_TX_OUT_OF_PLACE,
\
+.flags = AV_TX_UNALIGNED | AV_TX_INPLACE | FF_TX_OUT_OF_PLACE,
\
.factors= { N, TX_FACTOR_ANY },
\
.min_len= N*2,
\
.max_len= TX_LEN_UNLIMITED,
\
___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
To unsubscribe, visit link above, or email
ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] lavu/tx: generalize MDCTs
ffmpeg | branch: master | Lynne | Sat Sep 10 02:28:10 2022
+0200| [51172223fd1a5b71b46fc0d398f4fdc9ed081b83] | committer: Lynne
lavu/tx: generalize MDCTs
The same code can perform any-length MDCTs with minimal changes.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=51172223fd1a5b71b46fc0d398f4fdc9ed081b83
---
libavutil/tx_template.c | 75 ++---
1 file changed, 46 insertions(+), 29 deletions(-)
diff --git a/libavutil/tx_template.c b/libavutil/tx_template.c
index 542c15e480..1d4c4d294b 100644
--- a/libavutil/tx_template.c
+++ b/libavutil/tx_template.c
@@ -940,12 +940,12 @@ static const FFTXCodelet
TX_NAME(ff_tx_mdct_naive_inv_def) = {
.prio = FF_TX_PRIO_MIN,
};
-static av_cold int TX_NAME(ff_tx_mdct_sr_init)(AVTXContext *s,
- const FFTXCodelet *cd,
- uint64_t flags,
- FFTXCodeletOptions *opts,
- int len, int inv,
- const void *scale)
+static av_cold int TX_NAME(ff_tx_mdct_init)(AVTXContext *s,
+const FFTXCodelet *cd,
+uint64_t flags,
+FFTXCodeletOptions *opts,
+int len, int inv,
+const void *scale)
{
int ret;
FFTXCodeletOptions sub_opts = { .invert_lookup = inv };
@@ -955,32 +955,49 @@ static av_cold int
TX_NAME(ff_tx_mdct_sr_init)(AVTXContext *s,
flags &= ~FF_TX_OUT_OF_PLACE; /* We want the subtransform to be */
flags |= AV_TX_INPLACE; /* in-place */
-flags |= FF_TX_PRESHUFFLE; /* This function handles the permute step */
+flags |= FF_TX_PRESHUFFLE; /* First try with an in-place transform */
if ((ret = ff_tx_init_subtx(s, TX_TYPE(FFT), flags, &sub_opts, len >> 1,
-inv, scale)))
-return ret;
+inv, scale))) {
+flags &= ~FF_TX_PRESHUFFLE; /* Now try with a generic FFT */
+if ((ret = ff_tx_init_subtx(s, TX_TYPE(FFT), flags, &sub_opts, len >>
1,
+inv, scale)))
+return ret;
+}
+
+/* If we need to preshuffle just steal the map from the subcontext */
+if (s->sub[0].flags & FF_TX_PRESHUFFLE) {
+s->map = s->sub[0].map;
+s->sub[0].map = NULL;
+} else {
+s->map = av_malloc((len >> 1)*sizeof(*s->map));
+if (!s->map)
+return AVERROR(ENOMEM);
+
+for (int i = 0; i < len >> 1; i++)
+s->map[i] = i;
+}
-if ((ret = TX_TAB(ff_tx_mdct_gen_exp)(s, inv ? s->sub->map : NULL)))
+if ((ret = TX_TAB(ff_tx_mdct_gen_exp)(s, inv ? s->map : NULL)))
return ret;
/* Saves a multiply in a hot path. */
if (inv)
for (int i = 0; i < (s->len >> 1); i++)
-s->sub->map[i] <<= 1;
+s->map[i] <<= 1;
return 0;
}
-static void TX_NAME(ff_tx_mdct_sr_fwd)(AVTXContext *s, void *_dst, void *_src,
- ptrdiff_t stride)
+static void TX_NAME(ff_tx_mdct_fwd)(AVTXContext *s, void *_dst, void *_src,
+ptrdiff_t stride)
{
TXSample *src = _src, *dst = _dst;
TXComplex *exp = s->exp, tmp, *z = _dst;
const int len2 = s->len >> 1;
const int len4 = s->len >> 2;
const int len3 = len2 * 3;
-const int *sub_map = s->sub->map;
+const int *sub_map = s->map;
stride /= sizeof(*dst);
@@ -1011,14 +1028,14 @@ static void TX_NAME(ff_tx_mdct_sr_fwd)(AVTXContext *s,
void *_dst, void *_src,
}
}
-static void TX_NAME(ff_tx_mdct_sr_inv)(AVTXContext *s, void *_dst, void *_src,
- ptrdiff_t stride)
+static void TX_NAME(ff_tx_mdct_inv)(AVTXContext *s, void *_dst, void *_src,
+ptrdiff_t stride)
{
TXComplex *z = _dst, *exp = s->exp;
const TXSample *src = _src, *in1, *in2;
const int len2 = s->len >> 1;
const int len4 = s->len >> 2;
-const int *sub_map = s->sub->map;
+const int *sub_map = s->map;
stride /= sizeof(*src);
in1 = src;
@@ -1043,28 +1060,28 @@ static void TX_NAME(ff_tx_mdct_sr_inv)(AVTXContext *s,
void *_dst, void *_src,
}
}
-static const FFTXCodelet TX_NAME(ff_tx_mdct_sr_fwd_def) = {
-.name = TX_NAME_STR("mdct_sr_fwd"),
-.function = TX_NAME(ff_tx_mdct_sr_fwd),
+static const FFTXCodelet TX_NAME(ff_tx_mdct_fwd_def) = {
+.name = TX_NAME_STR("mdct_fwd"),
+.function = TX_NAME(ff_tx_mdct_fwd),
.type = TX_TYPE(MDCT),
.flags = AV_TX_UNALIGNED | FF_TX_OUT_OF_PLACE | FF_TX_FORWARD_ONLY,
-.factors[0] = 2,
+.factors= { 2, TX_FACTOR_A
[FFmpeg-cvslog] lavu/tx: propagate the codelet flags into the context
ffmpeg | branch: master | Lynne | Sat Sep 10 02:26:02 2022 +0200| [8c283e8fe631135a0c36d50f9c8d558f43cfef7b] | committer: Lynne lavu/tx: propagate the codelet flags into the context The field is documented as a combination of both. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8c283e8fe631135a0c36d50f9c8d558f43cfef7b --- libavutil/tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavutil/tx.c b/libavutil/tx.c index da8ebddd9a..aeb0d9dada 100644 --- a/libavutil/tx.c +++ b/libavutil/tx.c @@ -620,7 +620,7 @@ av_cold int ff_tx_init_subtx(AVTXContext *s, enum AVTXType type, sctx->len= len; sctx->inv= inv; sctx->type = type; -sctx->flags = flags; +sctx->flags = cd->flags | flags; sctx->cd_self= cd; s->fn[s->nb_sub] = cd->function; ___ ffmpeg-cvslog mailing list ffmpeg-cvslog@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog To unsubscribe, visit link above, or email ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-cvslog] lavu/tx: rotate 3 & 15-point exptabs
ffmpeg | branch: master | Lynne | Sat Sep 10 02:31:43 2022 +0200| [c92edd969aaf8b12434ff4bd731aa4bc5548fbbf] | committer: Lynne lavu/tx: rotate 3 & 15-point exptabs This just inverts their signs. Simplifies SIMD. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c92edd969aaf8b12434ff4bd731aa4bc5548fbbf --- libavutil/tx_template.c | 10 +- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/libavutil/tx_template.c b/libavutil/tx_template.c index 1d4c4d294b..0c7ddd26f6 100644 --- a/libavutil/tx_template.c +++ b/libavutil/tx_template.c @@ -109,11 +109,11 @@ static av_cold void TX_TAB(ff_tx_init_tab_53)(void) TX_TAB(ff_tx_tab_53)[0] = RESCALE(cos(2 * M_PI / 12)); TX_TAB(ff_tx_tab_53)[1] = RESCALE(cos(2 * M_PI / 12)); TX_TAB(ff_tx_tab_53)[2] = RESCALE(cos(2 * M_PI / 6)); -TX_TAB(ff_tx_tab_53)[3] = RESCALE(cos(2 * M_PI / 6)); +TX_TAB(ff_tx_tab_53)[3] = RESCALE(cos(8 * M_PI / 6)); TX_TAB(ff_tx_tab_53)[4] = RESCALE(cos(2 * M_PI / 5)); -TX_TAB(ff_tx_tab_53)[5] = RESCALE(sin(2 * M_PI / 5)); +TX_TAB(ff_tx_tab_53)[5] = RESCALE(sin(8 * M_PI / 5)); TX_TAB(ff_tx_tab_53)[6] = RESCALE(cos(2 * M_PI / 10)); -TX_TAB(ff_tx_tab_53)[7] = RESCALE(sin(2 * M_PI / 10)); +TX_TAB(ff_tx_tab_53)[7] = RESCALE(sin(6 * M_PI / 5)); } static av_cold void TX_TAB(ff_tx_init_tab_7)(void) @@ -226,8 +226,8 @@ static av_always_inline void NAME(TXComplex *out, TXComplex *in,\ \ SMUL(t[4].re, t[0].re, tab[4], tab[6], t[2].re, t[0].re); \ SMUL(t[4].im, t[0].im, tab[4], tab[6], t[2].im, t[0].im); \ -CMUL(t[5].re, t[1].re, tab[5], tab[7], t[3].re, t[1].re); \ -CMUL(t[5].im, t[1].im, tab[5], tab[7], t[3].im, t[1].im); \ +CMUL(t[5].re, t[1].re, -tab[5], -tab[7], t[3].re, t[1].re); \ +CMUL(t[5].im, t[1].im, -tab[5], -tab[7], t[3].im, t[1].im); \ \ BF(z0[0].re, z0[3].re, t[0].re, t[1].re); \ BF(z0[0].im, z0[3].im, t[0].im, t[1].im); \ ___ ffmpeg-cvslog mailing list ffmpeg-cvslog@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog To unsubscribe, visit link above, or email ffmpeg-cvslog-requ...@ffmpeg.org with subject "unsubscribe".
