# HG changeset patch # User michael # Date 1187654598 0 # Node ID 0074dd43ae5e07858b5392c81a21f763e2b4e8ad # Parent 225867be27f7c0a6b1d833f6a495134f33bc711c 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 diff -r 225867be27f7 -r 0074dd43ae5e i386/snowdsp_mmx.c --- a/i386/snowdsp_mmx.c Tue Aug 21 00:02:19 2007 +0000 +++ b/i386/snowdsp_mmx.c Tue Aug 21 00:03:18 2007 +0000 @@ -498,9 +498,8 @@ 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_sra("1","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_add(REG_a,"xmm0","xmm2","xmm4","xmm6") snow_vertical_compose_sse2_store(REG_a,"xmm0","xmm2","xmm4","xmm6")