| Message ID | 20220725111541.44618-1-hum@semihalf.com |
|---|---|
| State | New |
| Headers | show |
| Series | [FFmpeg-devel] lavc/aarch64: Add neon implementation for sse4 | expand |
| Context | Check | Description |
|---|---|---|
| yinshiyou/configure_loongarch64 | warning | Failed to apply patch |
| andriy/configure_x86 | warning | Failed to apply patch |
Your latest set of patches didn’t get interpreted correctly by the patchwork tool. I suspect it took them in the wrong order.
https://patchwork.ffmpeg.org/project/ffmpeg/list/?series=&submitter=1479&state=&q=&archive=&delegate=
There is one more place to move the sub, cmp and branch instructions apart in sse16_neon. It doesn't seem to make any difference to Neoverse N1 and V1 and it may help A53.
I didn't see anything else.
Thanks!
--
Jonathan Swinney
On 7/25/22, 6:16 AM, "Hubert Mazur" <hum@semihalf.com> wrote:
CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.
Provide neon implementation for sse4 function.
Performance comparison tests are shown below.
- sse_2_c: 74.0
- sse_2_neon: 24.0
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum@semihalf.com>
---
libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++
libavcodec/aarch64/me_cmp_neon.S | 65 ++++++++++++++++++++++++
2 files changed, 68 insertions(+)
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c
index 3ff5767bd0..72a2062e7e 100644
--- a/libavcodec/aarch64/me_cmp_init_aarch64.c
+++ b/libavcodec/aarch64/me_cmp_init_aarch64.c
@@ -32,6 +32,8 @@ int ff_pix_abs16_x2_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
ptrdiff_t stride, int h);
+int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
+ ptrdiff_t stride, int h);
av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
{
@@ -44,5 +46,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx)
c->sad[0] = ff_pix_abs16_neon;
c->sse[0] = sse16_neon;
+ c->sse[2] = sse4_neon;
}
}
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index 98c912b608..3336d88848 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -352,3 +352,68 @@ function sse16_neon, export=1
ret
endfunc
+
+function sse4_neon, export=1
+ // x0 - unused
+ // x1 - pix1
+ // x2 - pix2
+ // x3 - stride
+ // w4 - h
+
+ movi d18, #0
+ movi d17, #0
+ cmp w4, #4
+ b.le 2f
+
+// make 4 iterations at once
+1:
+
+ // res = abs(pix1[0] - pix2[0])
+ // res * res
+
+ ld1 {v0.s}[0], [x1], x3 // Load pix1, first iteration
+ ld1 {v1.s}[0], [x2], x3 // Load pix2, first iteration
+ uabdl v30.8h, v0.8b, v1.8b // Absolute difference, first iteration
+ ld1 {v2.s}[0], [x1], x3 // Load pix1, second iteration
+ ld1 {v3.s}[0], [x2], x3 // Load pix2, second iteration
+ umull v16.4s, v30.4h, v30.4h // Multiply vectors, first iteration
+ uabdl v29.8h, v2.8b, v3.8b // Absolute difference, second iteration
+ ld1 {v4.s}[0], [x1], x3 // Load pix1, third iteration
+ ld1 {v5.s}[0], [x2], x3 // Load pix2, third iteration
+ umlal v16.4s, v29.4h, v29.4h // Multiply and accumulate, second iteration
+ uabdl v28.8h, v4.8b, v5.8b // Absolute difference, third iteration
+ ld1 {v6.s}[0], [x1], x3 // Load pix1, fourth iteration
+ ld1 {v7.s}[0], [x2], x3 // Load pix2, fourth iteration
+ umlal v16.4s, v28.4h, v28.4h // Multiply and accumulate, third iteration
+ uabdl v27.8h, v6.8b, v7.8b // Absolue difference, fourth iteration
+ umlal v16.4s, v27.4h, v27.4h // Multiply and accumulate, fourth iteration
+
+ uaddlv d17, v16.4s // Add vector
+ add d18, d18, d17
+
+ sub w4, w4, #4
+ cmp w4, #4
+ b.ge 1b
+
+ cbnz w4, 2f
+ fmov w0, s18
+
+ ret
+
+// iterate by one
+2:
+ ld1 {v0.s}[0], [x1], x3 // Load pix1
+ ld1 {v1.s}[0], [x2], x3 // Load pix2
+ uabdl v30.8h, v0.8b, v1.8b
+ umull v16.4s, v30.4h, v30.4h
+
+ uaddlv d17, v16.4s
+ add d18, d18, d17
+
+ subs w4, w4, #1
+ b.ne 2b
+ fmov w0, s18
+
+ ret
+
+endfunc
--
2.34.1
> There is one more place to move the sub, cmp and branch instructions apart in sse16_neon. It doesn't seem to make any difference to Neoverse N1 and V1 and it may help A53. Sorry-- I meant sse4_neon. -- Jonathan Swinney On 7/28/22, 1:50 PM, "Swinney, Jonathan" <jswinney@amazon.com> wrote: Your latest set of patches didn’t get interpreted correctly by the patchwork tool. I suspect it took them in the wrong order. https://patchwork.ffmpeg.org/project/ffmpeg/list/?series=&submitter=1479&state=&q=&archive=&delegate= There is one more place to move the sub, cmp and branch instructions apart in sse16_neon. It doesn't seem to make any difference to Neoverse N1 and V1 and it may help A53. I didn't see anything else. Thanks! -- Jonathan Swinney On 7/25/22, 6:16 AM, "Hubert Mazur" <hum@semihalf.com> wrote: CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe. Provide neon implementation for sse4 function. Performance comparison tests are shown below. - sse_2_c: 74.0 - sse_2_neon: 24.0 Benchmarks and tests are run with checkasm tool on AWS Graviton 3. Signed-off-by: Hubert Mazur <hum@semihalf.com> --- libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++ libavcodec/aarch64/me_cmp_neon.S | 65 ++++++++++++++++++++++++ 2 files changed, 68 insertions(+) diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c index 3ff5767bd0..72a2062e7e 100644 --- a/libavcodec/aarch64/me_cmp_init_aarch64.c +++ b/libavcodec/aarch64/me_cmp_init_aarch64.c @@ -32,6 +32,8 @@ int ff_pix_abs16_x2_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h); +int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, + ptrdiff_t stride, int h); av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) { @@ -44,5 +46,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) c->sad[0] = ff_pix_abs16_neon; c->sse[0] = sse16_neon; + c->sse[2] = sse4_neon; } } diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S index 98c912b608..3336d88848 100644 --- a/libavcodec/aarch64/me_cmp_neon.S +++ b/libavcodec/aarch64/me_cmp_neon.S @@ -352,3 +352,68 @@ function sse16_neon, export=1 ret endfunc + +function sse4_neon, export=1 + // x0 - unused + // x1 - pix1 + // x2 - pix2 + // x3 - stride + // w4 - h + + movi d18, #0 + movi d17, #0 + cmp w4, #4 + b.le 2f + +// make 4 iterations at once +1: + + // res = abs(pix1[0] - pix2[0]) + // res * res + + ld1 {v0.s}[0], [x1], x3 // Load pix1, first iteration + ld1 {v1.s}[0], [x2], x3 // Load pix2, first iteration + uabdl v30.8h, v0.8b, v1.8b // Absolute difference, first iteration + ld1 {v2.s}[0], [x1], x3 // Load pix1, second iteration + ld1 {v3.s}[0], [x2], x3 // Load pix2, second iteration + umull v16.4s, v30.4h, v30.4h // Multiply vectors, first iteration + uabdl v29.8h, v2.8b, v3.8b // Absolute difference, second iteration + ld1 {v4.s}[0], [x1], x3 // Load pix1, third iteration + ld1 {v5.s}[0], [x2], x3 // Load pix2, third iteration + umlal v16.4s, v29.4h, v29.4h // Multiply and accumulate, second iteration + uabdl v28.8h, v4.8b, v5.8b // Absolute difference, third iteration + ld1 {v6.s}[0], [x1], x3 // Load pix1, fourth iteration + ld1 {v7.s}[0], [x2], x3 // Load pix2, fourth iteration + umlal v16.4s, v28.4h, v28.4h // Multiply and accumulate, third iteration + uabdl v27.8h, v6.8b, v7.8b // Absolue difference, fourth iteration + umlal v16.4s, v27.4h, v27.4h // Multiply and accumulate, fourth iteration + + uaddlv d17, v16.4s // Add vector + add d18, d18, d17 + + sub w4, w4, #4 + cmp w4, #4 + b.ge 1b + + cbnz w4, 2f + fmov w0, s18 + + ret + +// iterate by one +2: + ld1 {v0.s}[0], [x1], x3 // Load pix1 + ld1 {v1.s}[0], [x2], x3 // Load pix2 + uabdl v30.8h, v0.8b, v1.8b + umull v16.4s, v30.4h, v30.4h + + uaddlv d17, v16.4s + add d18, d18, d17 + + subs w4, w4, #1 + b.ne 2b + fmov w0, s18 + + ret + +endfunc -- 2.34.1
Yes, it seems that they are misplaced or each is treated as a new series and thus can't be applied. I will send the whole batch again after the first review, so some issues could be fixed. Thanks for the feedback! On Thu, Jul 28, 2022 at 8:51 PM Swinney, Jonathan <jswinney@amazon.com> wrote: > > There is one more place to move the sub, cmp and branch instructions > apart in sse16_neon. It doesn't seem to make any difference to Neoverse N1 > and V1 and it may help A53. > > Sorry-- I meant sse4_neon. > > -- > > Jonathan Swinney > > On 7/28/22, 1:50 PM, "Swinney, Jonathan" <jswinney@amazon.com> wrote: > > Your latest set of patches didn’t get interpreted correctly by the > patchwork tool. I suspect it took them in the wrong order. > > > https://patchwork.ffmpeg.org/project/ffmpeg/list/?series=&submitter=1479&state=&q=&archive=&delegate= > > There is one more place to move the sub, cmp and branch instructions > apart in sse16_neon. It doesn't seem to make any difference to Neoverse N1 > and V1 and it may help A53. > > I didn't see anything else. > > Thanks! > -- > > Jonathan Swinney > > On 7/25/22, 6:16 AM, "Hubert Mazur" <hum@semihalf.com> wrote: > > CAUTION: This email originated from outside of the organization. > Do not click links or open attachments unless you can confirm the sender > and know the content is safe. > > > > Provide neon implementation for sse4 function. > > Performance comparison tests are shown below. > - sse_2_c: 74.0 > - sse_2_neon: 24.0 > > Benchmarks and tests are run with checkasm tool on AWS Graviton 3. > > Signed-off-by: Hubert Mazur <hum@semihalf.com> > --- > libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++ > libavcodec/aarch64/me_cmp_neon.S | 65 > ++++++++++++++++++++++++ > 2 files changed, 68 insertions(+) > > diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c > b/libavcodec/aarch64/me_cmp_init_aarch64.c > index 3ff5767bd0..72a2062e7e 100644 > --- a/libavcodec/aarch64/me_cmp_init_aarch64.c > +++ b/libavcodec/aarch64/me_cmp_init_aarch64.c > @@ -32,6 +32,8 @@ int ff_pix_abs16_x2_neon(MpegEncContext *v, > uint8_t *pix1, uint8_t *pix2, > > int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, > ptrdiff_t stride, int h); > +int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, > + ptrdiff_t stride, int h); > > av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, > AVCodecContext *avctx) > { > @@ -44,5 +46,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext > *c, AVCodecContext *avctx) > > c->sad[0] = ff_pix_abs16_neon; > c->sse[0] = sse16_neon; > + c->sse[2] = sse4_neon; > } > } > diff --git a/libavcodec/aarch64/me_cmp_neon.S > b/libavcodec/aarch64/me_cmp_neon.S > index 98c912b608..3336d88848 100644 > --- a/libavcodec/aarch64/me_cmp_neon.S > +++ b/libavcodec/aarch64/me_cmp_neon.S > @@ -352,3 +352,68 @@ function sse16_neon, export=1 > ret > > endfunc > + > +function sse4_neon, export=1 > + // x0 - unused > + // x1 - pix1 > + // x2 - pix2 > + // x3 - stride > + // w4 - h > + > + movi d18, #0 > + movi d17, #0 > + cmp w4, #4 > + b.le 2f > + > +// make 4 iterations at once > +1: > + > + // res = abs(pix1[0] - pix2[0]) > + // res * res > + > + ld1 {v0.s}[0], [x1], x3 // Load > pix1, first iteration > + ld1 {v1.s}[0], [x2], x3 // Load > pix2, first iteration > + uabdl v30.8h, v0.8b, v1.8b // > Absolute difference, first iteration > + ld1 {v2.s}[0], [x1], x3 // Load > pix1, second iteration > + ld1 {v3.s}[0], [x2], x3 // Load > pix2, second iteration > + umull v16.4s, v30.4h, v30.4h // > Multiply vectors, first iteration > + uabdl v29.8h, v2.8b, v3.8b // > Absolute difference, second iteration > + ld1 {v4.s}[0], [x1], x3 // Load > pix1, third iteration > + ld1 {v5.s}[0], [x2], x3 // Load > pix2, third iteration > + umlal v16.4s, v29.4h, v29.4h // > Multiply and accumulate, second iteration > + uabdl v28.8h, v4.8b, v5.8b // > Absolute difference, third iteration > + ld1 {v6.s}[0], [x1], x3 // Load > pix1, fourth iteration > + ld1 {v7.s}[0], [x2], x3 // Load > pix2, fourth iteration > + umlal v16.4s, v28.4h, v28.4h // > Multiply and accumulate, third iteration > + uabdl v27.8h, v6.8b, v7.8b // > Absolue difference, fourth iteration > + umlal v16.4s, v27.4h, v27.4h // > Multiply and accumulate, fourth iteration > + > + uaddlv d17, v16.4s // Add > vector > + add d18, d18, d17 > + > + sub w4, w4, #4 > + cmp w4, #4 > + b.ge 1b > + > + cbnz w4, 2f > + fmov w0, s18 > + > + ret > + > +// iterate by one > +2: > + ld1 {v0.s}[0], [x1], x3 // Load > pix1 > + ld1 {v1.s}[0], [x2], x3 // Load > pix2 > + uabdl v30.8h, v0.8b, v1.8b > + umull v16.4s, v30.4h, v30.4h > + > + uaddlv d17, v16.4s > + add d18, d18, d17 > + > + subs w4, w4, #1 > + b.ne 2b > + fmov w0, s18 > + > + ret > + > +endfunc > -- > 2.34.1 > > > >
On Mon, 25 Jul 2022, Hubert Mazur wrote: > Provide neon implementation for sse4 function. > > Performance comparison tests are shown below. > - sse_2_c: 74.0 > - sse_2_neon: 24.0 > > Benchmarks and tests are run with checkasm tool on AWS Graviton 3. > > Signed-off-by: Hubert Mazur <hum@semihalf.com> > --- > libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++ > libavcodec/aarch64/me_cmp_neon.S | 65 ++++++++++++++++++++++++ > 2 files changed, 68 insertions(+) > > diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c > index 3ff5767bd0..72a2062e7e 100644 > --- a/libavcodec/aarch64/me_cmp_init_aarch64.c > +++ b/libavcodec/aarch64/me_cmp_init_aarch64.c > @@ -32,6 +32,8 @@ int ff_pix_abs16_x2_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, > > int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, > ptrdiff_t stride, int h); > +int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, > + ptrdiff_t stride, int h); > > av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) > { > @@ -44,5 +46,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) > > c->sad[0] = ff_pix_abs16_neon; > c->sse[0] = sse16_neon; > + c->sse[2] = sse4_neon; > } > } > diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S > index 98c912b608..3336d88848 100644 > --- a/libavcodec/aarch64/me_cmp_neon.S > +++ b/libavcodec/aarch64/me_cmp_neon.S > @@ -352,3 +352,68 @@ function sse16_neon, export=1 > ret > > endfunc > + > +function sse4_neon, export=1 > + // x0 - unused > + // x1 - pix1 > + // x2 - pix2 > + // x3 - stride > + // w4 - h > + > + movi d18, #0 > + movi d17, #0 In the current implementation, it doesn't seem like d17 needs to be initialized here > + cmp w4, #4 > + b.le 2f > + > +// make 4 iterations at once > +1: > + > + // res = abs(pix1[0] - pix2[0]) > + // res * res > + > + ld1 {v0.s}[0], [x1], x3 // Load pix1, first iteration > + ld1 {v1.s}[0], [x2], x3 // Load pix2, first iteration > + uabdl v30.8h, v0.8b, v1.8b // Absolute difference, first iteration Right now, half of the values calculated by uabdl are unused; you could try loading two iterations into v0.s[0] and v0.s[1] so that the full .8b register gets used. Doing that would reduce the number of uabdl instructions from 4 to 2 - but it might make it harder to interleave instructions efficiently. So after all, maybe it's not worth if, it we can make the loads more efficiently interleaved this way? Again, also here, it'd be good to interleave things more efficiently, e.g. like this: ld1 first ld1 first ld1 second ld1 second uabdl first ld1 third ld1 third uabdl second umull first ld1 fourth ld1 fourth uabdl third umlal second uabdl fourth umlal third umlal fourth > + ld1 {v2.s}[0], [x1], x3 // Load pix1, second iteration > + ld1 {v3.s}[0], [x2], x3 // Load pix2, second iteration > + umull v16.4s, v30.4h, v30.4h // Multiply vectors, first iteration > + uabdl v29.8h, v2.8b, v3.8b // Absolute difference, second iteration > + ld1 {v4.s}[0], [x1], x3 // Load pix1, third iteration > + ld1 {v5.s}[0], [x2], x3 // Load pix2, third iteration > + umlal v16.4s, v29.4h, v29.4h // Multiply and accumulate, second iteration > + uabdl v28.8h, v4.8b, v5.8b // Absolute difference, third iteration > + ld1 {v6.s}[0], [x1], x3 // Load pix1, fourth iteration > + ld1 {v7.s}[0], [x2], x3 // Load pix2, fourth iteration > + umlal v16.4s, v28.4h, v28.4h // Multiply and accumulate, third iteration > + uabdl v27.8h, v6.8b, v7.8b // Absolue difference, fourth iteration > + umlal v16.4s, v27.4h, v27.4h // Multiply and accumulate, fourth iteration > + > + uaddlv d17, v16.4s // Add vector > + add d18, d18, d17 As usual, don't do any *add*v within the loop, defer it as far as possible. Here you're accumulating in 32 bit elements, so it will surely fit the results from the whole algorithm. Also, if you get rid of the uaddlv here, you can also accumulate into two separate .4s registers that you only add at the end; that allows two umlal instructions to possibly execute in parallel without waiting for each other (provided that the cpu has enough execution units for that). // Martin
diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c index 3ff5767bd0..72a2062e7e 100644 --- a/libavcodec/aarch64/me_cmp_init_aarch64.c +++ b/libavcodec/aarch64/me_cmp_init_aarch64.c @@ -32,6 +32,8 @@ int ff_pix_abs16_x2_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h); +int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, + ptrdiff_t stride, int h); av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) { @@ -44,5 +46,6 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) c->sad[0] = ff_pix_abs16_neon; c->sse[0] = sse16_neon; + c->sse[2] = sse4_neon; } } diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S index 98c912b608..3336d88848 100644 --- a/libavcodec/aarch64/me_cmp_neon.S +++ b/libavcodec/aarch64/me_cmp_neon.S @@ -352,3 +352,68 @@ function sse16_neon, export=1 ret endfunc + +function sse4_neon, export=1 + // x0 - unused + // x1 - pix1 + // x2 - pix2 + // x3 - stride + // w4 - h + + movi d18, #0 + movi d17, #0 + cmp w4, #4 + b.le 2f + +// make 4 iterations at once +1: + + // res = abs(pix1[0] - pix2[0]) + // res * res + + ld1 {v0.s}[0], [x1], x3 // Load pix1, first iteration + ld1 {v1.s}[0], [x2], x3 // Load pix2, first iteration + uabdl v30.8h, v0.8b, v1.8b // Absolute difference, first iteration + ld1 {v2.s}[0], [x1], x3 // Load pix1, second iteration + ld1 {v3.s}[0], [x2], x3 // Load pix2, second iteration + umull v16.4s, v30.4h, v30.4h // Multiply vectors, first iteration + uabdl v29.8h, v2.8b, v3.8b // Absolute difference, second iteration + ld1 {v4.s}[0], [x1], x3 // Load pix1, third iteration + ld1 {v5.s}[0], [x2], x3 // Load pix2, third iteration + umlal v16.4s, v29.4h, v29.4h // Multiply and accumulate, second iteration + uabdl v28.8h, v4.8b, v5.8b // Absolute difference, third iteration + ld1 {v6.s}[0], [x1], x3 // Load pix1, fourth iteration + ld1 {v7.s}[0], [x2], x3 // Load pix2, fourth iteration + umlal v16.4s, v28.4h, v28.4h // Multiply and accumulate, third iteration + uabdl v27.8h, v6.8b, v7.8b // Absolue difference, fourth iteration + umlal v16.4s, v27.4h, v27.4h // Multiply and accumulate, fourth iteration + + uaddlv d17, v16.4s // Add vector + add d18, d18, d17 + + sub w4, w4, #4 + cmp w4, #4 + b.ge 1b + + cbnz w4, 2f + fmov w0, s18 + + ret + +// iterate by one +2: + ld1 {v0.s}[0], [x1], x3 // Load pix1 + ld1 {v1.s}[0], [x2], x3 // Load pix2 + uabdl v30.8h, v0.8b, v1.8b + umull v16.4s, v30.4h, v30.4h + + uaddlv d17, v16.4s + add d18, d18, d17 + + subs w4, w4, #1 + b.ne 2b + fmov w0, s18 + + ret + +endfunc
Provide neon implementation for sse4 function. Performance comparison tests are shown below. - sse_2_c: 74.0 - sse_2_neon: 24.0 Benchmarks and tests are run with checkasm tool on AWS Graviton 3. Signed-off-by: Hubert Mazur <hum@semihalf.com> --- libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++ libavcodec/aarch64/me_cmp_neon.S | 65 ++++++++++++++++++++++++ 2 files changed, 68 insertions(+)