[x265] [PATCH] asm: avx2 code for sse_ss[32x32] for 10 bpp (1935 -> 1030)

sumalatha at multicorewareinc.com sumalatha at multicorewareinc.com
Thu May 21 08:33:05 CEST 2015


# HG changeset patch
# User Sumalatha Polureddy
# Date 1432189730 -19800
#      Thu May 21 11:58:50 2015 +0530
# Node ID 184d645ee5f097dfb94d139cd23c3f1e3feddec1
# Parent  dc4fcfc574ade14ecc841797ad08be9753fad58e
asm: avx2 code for sse_ss[32x32] for 10 bpp (1935 -> 1030)

sse2:
sse_ss[32x32]  4.71x    1935.46         9113.11

avx2:
sse_ss[32x32]  8.85x    1030.08         9112.05

diff -r dc4fcfc574ad -r 184d645ee5f0 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Wed May 20 12:17:44 2015 -0500
+++ b/source/common/x86/asm-primitives.cpp	Thu May 21 11:58:50 2015 +0530
@@ -1259,6 +1259,7 @@
 
         p.cu[BLOCK_32x32].ssd_s = x265_pixel_ssd_s_32_avx2;
         p.cu[BLOCK_16x16].sse_ss = x265_pixel_ssd_ss_16x16_avx2;
+        p.cu[BLOCK_32x32].sse_ss = x265_pixel_ssd_ss_32x32_avx2;
 
         p.quant = x265_quant_avx2;
         p.nquant = x265_nquant_avx2;
diff -r dc4fcfc574ad -r 184d645ee5f0 source/common/x86/ssd-a.asm
--- a/source/common/x86/ssd-a.asm	Wed May 20 12:17:44 2015 -0500
+++ b/source/common/x86/ssd-a.asm	Thu May 21 11:58:50 2015 +0530
@@ -347,6 +347,7 @@
 INIT_YMM avx2
 SSD_ONE    16,  8
 SSD_ONE    16, 16
+SSD_ONE    32, 32
 %endif ; HIGH_BIT_DEPTH
 
 ;-----------------------------------------------------------------------------


More information about the x265-devel mailing list