From 355903f5a8ba254b439903a77d75f1d9c0d3f092 Mon Sep 17 00:00:00 2001
From: Michael Niedermayer <michaelni@gmx.at>
Date: Tue, 22 Aug 2006 11:11:04 +0000
Subject: [PATCH] make high precission synth filter 3 times faster on x86

Originally committed as revision 6046 to svn://svn.ffmpeg.org/ffmpeg/trunk
---
 libavcodec/mpegaudiodec.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/libavcodec/mpegaudiodec.c b/libavcodec/mpegaudiodec.c
index 53a3890..9593e03 100644
--- a/libavcodec/mpegaudiodec.c
+++ b/libavcodec/mpegaudiodec.c
@@ -784,8 +784,13 @@ static inline int round_sample(int64_t *sum)
     return sum1;
 }
 
+#ifdef ARCH_X86
+/* ask gcc devels why this is 3 times faster then the generic code below */
+#define MULS(ra, rb) \
+    ({ int64_t rt; asm ("imull %2\n\t" : "=A"(rt) : "a" (ra), "g" (rb)); rt; })
+#else
 #define MULS(ra, rb) MUL64(ra, rb)
-
+#endif
 #endif
 
 #define SUM8(sum, op, w, p) \
-- 
2.7.4