From 1ffbbef2177f4b6e633f6d52d364c20a17e3410f Mon Sep 17 00:00:00 2001
From: Michael Niedermayer <michaelni@gmx.at>
Date: Mon, 20 Aug 2007 23:59:20 +0000
Subject: [PATCH] Simplify and speedup code, reduce needed headroom by 2 bits
in the 3rd vertical lifting step of the SSE2 code (untested)
Originally committed as revision 10163 to svn://svn.ffmpeg.org/ffmpeg/trunk
---
libavcodec/i386/snowdsp_mmx.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/libavcodec/i386/snowdsp_mmx.c b/libavcodec/i386/snowdsp_mmx.c
index 7b3214dbb49..229c7218224 100644
--- a/libavcodec/i386/snowdsp_mmx.c
+++ b/libavcodec/i386/snowdsp_mmx.c
@@ -483,18 +483,17 @@ void ff_snow_vertical_compose97i_sse2(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, DWT
snow_vertical_compose_sse2_sub("xmm1","xmm3","xmm5","xmm7","xmm0","xmm2","xmm4","xmm6")
snow_vertical_compose_sse2_store(REG_S,"xmm0","xmm2","xmm4","xmm6")
"mov %2, %%"REG_a" \n\t"
- snow_vertical_compose_sse2_load(REG_c,"xmm1","xmm3","xmm5","xmm7")
snow_vertical_compose_sse2_add(REG_a,"xmm0","xmm2","xmm4","xmm6")
- snow_vertical_compose_sse2_sll("2","xmm1","xmm3","xmm5","xmm7")\
- snow_vertical_compose_sse2_r2r_add("xmm1","xmm3","xmm5","xmm7","xmm0","xmm2","xmm4","xmm6")
+ snow_vertical_compose_sse2_sra("2","xmm0","xmm2","xmm4","xmm6")
+ snow_vertical_compose_sse2_add(REG_c,"xmm0","xmm2","xmm4","xmm6")
"pcmpeqd %%xmm1, %%xmm1 \n\t"
"pslld $31, %%xmm1 \n\t"
- "psrld $28, %%xmm1 \n\t"
+ "psrld $30, %%xmm1 \n\t"
"mov %1, %%"REG_S" \n\t"
snow_vertical_compose_sse2_r2r_add("xmm1","xmm1","xmm1","xmm1","xmm0","xmm2","xmm4","xmm6")
- snow_vertical_compose_sse2_sra("4","xmm0","xmm2","xmm4","xmm6")
+ snow_vertical_compose_sse2_sra("2","xmm0","xmm2","xmm4","xmm6")
snow_vertical_compose_sse2_add(REG_c,"xmm0","xmm2","xmm4","xmm6")
snow_vertical_compose_sse2_store(REG_c,"xmm0","xmm2","xmm4","xmm6")
snow_vertical_compose_sse2_add(REG_S,"xmm0","xmm2","xmm4","xmm6")
--
GitLab