Simplify and optimize the 4th vertical lifting step of the SSE2 code (untested)

This also reduces the needed headroom in that step by 1 bit

Originally committed as revision 10165 to svn://svn.ffmpeg.org/ffmpeg/trunk

diff --git a/libavcodec/i386/snowdsp_mmx.c b/libavcodec/i386/snowdsp_mmx.c
index cfdfb95..b6509aa 100644 (file)
--- a/libavcodec/i386/snowdsp_mmx.c
+++ b/libavcodec/i386/snowdsp_mmx.c
@@ -498,9 +498,8 @@ void ff_snow_vertical_compose97i_sse2(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, DWT
         snow_vertical_compose_sse2_store(REG_c,"xmm0","xmm2","xmm4","xmm6")
         snow_vertical_compose_sse2_add(REG_S,"xmm0","xmm2","xmm4","xmm6")
         snow_vertical_compose_sse2_move("xmm0","xmm2","xmm4","xmm6","xmm1","xmm3","xmm5","xmm7")
-        snow_vertical_compose_sse2_r2r_add("xmm0","xmm2","xmm4","xmm6","xmm0","xmm2","xmm4","xmm6")
-        snow_vertical_compose_sse2_r2r_add("xmm1","xmm3","xmm5","xmm7","xmm0","xmm2","xmm4","xmm6")
         snow_vertical_compose_sse2_sra("1","xmm0","xmm2","xmm4","xmm6")
+        snow_vertical_compose_sse2_r2r_add("xmm1","xmm3","xmm5","xmm7","xmm0","xmm2","xmm4","xmm6")
         snow_vertical_compose_sse2_add(REG_a,"xmm0","xmm2","xmm4","xmm6")
         snow_vertical_compose_sse2_store(REG_a,"xmm0","xmm2","xmm4","xmm6")