From: "Martin Storsjö" <martin@martin.st>
To: ffmpeg-devel@ffmpeg.org
Cc: "Grzegorz Bernacki" <gjb@semihalf.com>,
"Jonathan Swinney" <jswinney@amazon.com>,
"Hubert Mazur" <hum@semihalf.com>,
"Martin Storsjö" <martin@martin.st>
Subject: [FFmpeg-devel] [PATCH 1/2] aarch64: me_cmp: Avoid redundant loads in ff_pix_abs16_y2_neon
Date: Wed, 28 Sep 2022 12:13:33 +0300
Message-ID: <20220928091334.7838-1-martin@martin.st> (raw)
This avoids one redundant load per row; pix3 from the previous
iteration can be used as pix2 in the next one.
Before: Cortex A53 A72 A73
pix_abs_0_2_neon: 138.0 59.7 48.0
After:
pix_abs_0_2_neon: 109.7 50.2 39.5
Signed-off-by: Martin Storsjö <martin@martin.st>
---
libavcodec/aarch64/me_cmp_neon.S | 24 ++++++++++--------------
1 file changed, 10 insertions(+), 14 deletions(-)
diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S
index 11af4849f9..832a7cb22d 100644
--- a/libavcodec/aarch64/me_cmp_neon.S
+++ b/libavcodec/aarch64/me_cmp_neon.S
@@ -326,9 +326,9 @@ function ff_pix_abs16_y2_neon, export=1
// w4 int h
// initialize buffers
+ ld1 {v1.16b}, [x2], x3 // Load pix2
movi v29.8h, #0 // clear the accumulator
movi v28.8h, #0 // clear the accumulator
- add x5, x2, x3 // pix2 + stride
cmp w4, #4
b.lt 2f
@@ -339,29 +339,25 @@ function ff_pix_abs16_y2_neon, export=1
// avg2(a, b) = (((a) + (b) + 1) >> 1)
// abs(x) = (x < 0 ? (-x) : (x))
- ld1 {v1.16b}, [x2], x3 // Load pix2 for first iteration
- ld1 {v2.16b}, [x5], x3 // Load pix3 for first iteration
+ ld1 {v2.16b}, [x2], x3 // Load pix3 for first iteration
ld1 {v0.16b}, [x1], x3 // Load pix1 for first iteration
urhadd v30.16b, v1.16b, v2.16b // Rounding halving add, first iteration
- ld1 {v4.16b}, [x2], x3 // Load pix2 for second iteration
- ld1 {v5.16b}, [x5], x3 // Load pix3 for second iteartion
+ ld1 {v5.16b}, [x2], x3 // Load pix3 for second iteartion
uabal v29.8h, v0.8b, v30.8b // Absolute difference of lower half, first iteration
uabal2 v28.8h, v0.16b, v30.16b // Absolute difference of upper half, first iteration
ld1 {v3.16b}, [x1], x3 // Load pix1 for second iteration
- urhadd v27.16b, v4.16b, v5.16b // Rounding halving add, second iteration
- ld1 {v7.16b}, [x2], x3 // Load pix2 for third iteration
- ld1 {v20.16b}, [x5], x3 // Load pix3 for third iteration
+ urhadd v27.16b, v2.16b, v5.16b // Rounding halving add, second iteration
+ ld1 {v20.16b}, [x2], x3 // Load pix3 for third iteration
uabal v29.8h, v3.8b, v27.8b // Absolute difference of lower half for second iteration
uabal2 v28.8h, v3.16b, v27.16b // Absolute difference of upper half for second iteration
ld1 {v6.16b}, [x1], x3 // Load pix1 for third iteration
- urhadd v26.16b, v7.16b, v20.16b // Rounding halving add, third iteration
- ld1 {v22.16b}, [x2], x3 // Load pix2 for fourth iteration
- ld1 {v23.16b}, [x5], x3 // Load pix3 for fourth iteration
+ urhadd v26.16b, v5.16b, v20.16b // Rounding halving add, third iteration
+ ld1 {v1.16b}, [x2], x3 // Load pix3 for fourth iteration
uabal v29.8h, v6.8b, v26.8b // Absolute difference of lower half for third iteration
uabal2 v28.8h, v6.16b, v26.16b // Absolute difference of upper half for third iteration
ld1 {v21.16b}, [x1], x3 // Load pix1 for fourth iteration
sub w4, w4, #4 // h-= 4
- urhadd v25.16b, v22.16b, v23.16b // Rounding halving add
+ urhadd v25.16b, v20.16b, v1.16b // Rounding halving add
cmp w4, #4
uabal v29.8h, v21.8b, v25.8b // Absolute difference of lower half for fourth iteration
uabal2 v28.8h, v21.16b, v25.16b // Absolute difference of upper half for fourth iteration
@@ -372,11 +368,11 @@ function ff_pix_abs16_y2_neon, export=1
// iterate by one
2:
- ld1 {v1.16b}, [x2], x3 // Load pix2
- ld1 {v2.16b}, [x5], x3 // Load pix3
+ ld1 {v2.16b}, [x2], x3 // Load pix3
subs w4, w4, #1
ld1 {v0.16b}, [x1], x3 // Load pix1
urhadd v30.16b, v1.16b, v2.16b // Rounding halving add
+ mov v1.16b, v2.16b // Shift pix3->pix2
uabal v29.8h, v30.8b, v0.8b
uabal2 v28.8h, v30.16b, v0.16b
--
2.25.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
next reply other threads:[~2022-09-28 9:13 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-28 9:13 Martin Storsjö [this message]
2022-09-28 9:13 ` [FFmpeg-devel] [PATCH 2/2] aarch64: me_cmp: Avoid using the non-unrolled codepath for the minimum unroll size Martin Storsjö
2022-09-28 9:19 ` Hubert Mazur
2022-09-28 9:18 ` [FFmpeg-devel] [PATCH 1/2] aarch64: me_cmp: Avoid redundant loads in ff_pix_abs16_y2_neon Hubert Mazur
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220928091334.7838-1-martin@martin.st \
--to=martin@martin.st \
--cc=ffmpeg-devel@ffmpeg.org \
--cc=gjb@semihalf.com \
--cc=hum@semihalf.com \
--cc=jswinney@amazon.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git