Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
* [FFmpeg-devel] [PATCH] lavc/mathops: simplify mid_pred() (PR #21188)
@ 2025-12-14  9:13 Rémi Denis-Courmont via ffmpeg-devel
  0 siblings, 0 replies; only message in thread
From: Rémi Denis-Courmont via ffmpeg-devel @ 2025-12-14  9:13 UTC (permalink / raw)
  To: ffmpeg-devel; +Cc: Rémi Denis-Courmont

PR #21188 opened by Rémi Denis-Courmont (Courmisch)
URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21188
Patch URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21188.patch

This reduces mid_pred() (i.e. median of 3) down to the minimum:
- 3 comparisons and 4 conditional moves, or
- 4 min/max,
whilst eliminating all branches.

The same algorithm is already implemented via inline assembler for some
architectures, but notably not Arm64 and RVA22. Besides, using C code
allows the compiler to schedule instruction properly.

See also https://godbolt.org/z/YKbabMG74 for comparisons.


From aa6c8d167f43c17fd7a324125f7bf668aec6d08a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?R=C3=A9mi=20Denis-Courmont?= <remi@remlab.net>
Date: Sun, 14 Dec 2025 11:06:13 +0200
Subject: [PATCH] lavc/mathops: simplify mid_pred()

This reduces mid_pred() (i.e. median of 3) down to the minimum:
- 3 comparisons and 4 conditional moves, or
- 4 min/max,
whilst eliminating all branches.

The same algorithm is already implemented via inline assembler for some
architectures, but notably not Arm64 and RVA22. Besides, using C code
allows the compiler to schedule instruction properly.
---
 libavcodec/mathops.h | 29 +++++++++++++++--------------
 1 file changed, 15 insertions(+), 14 deletions(-)

diff --git a/libavcodec/mathops.h b/libavcodec/mathops.h
index aa0bdfe956..4411d138b4 100644
--- a/libavcodec/mathops.h
+++ b/libavcodec/mathops.h
@@ -93,23 +93,24 @@ static av_always_inline unsigned UMULH(unsigned a, unsigned b){
 #endif
 
 /* median of 3 */
-#ifndef mid_pred
-#define mid_pred mid_pred
-static inline av_const int mid_pred(int a, int b, int c)
+static inline av_const int median3_c(int a, int b, int c)
 {
-    if(a>b){
-        if(c>b){
-            if(c>a) b=a;
-            else    b=c;
-        }
-    }else{
-        if(b>c){
-            if(c>a) b=c;
-            else    b=a;
-        }
+    int max2, min2, m;
+
+    if (a >= b) {
+        max2 = a;
+        min2 = b;
+    } else {
+        max2 = b;
+        min2 = a;
     }
-    return b;
+    m = (c >= max2) ? max2 : c;
+
+    return (m >= min2) ? m : min2;
 }
+
+#ifndef mid_pred
+#define mid_pred median3_c
 #endif
 
 #ifndef median4
-- 
2.49.1

_______________________________________________
ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org
To unsubscribe send an email to ffmpeg-devel-leave@ffmpeg.org

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2025-12-14  9:14 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-12-14  9:13 [FFmpeg-devel] [PATCH] lavc/mathops: simplify mid_pred() (PR #21188) Rémi Denis-Courmont via ffmpeg-devel

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