[x265] [PATCH 5 of 6] avx2:'integral32v' asm code-> 7.93x faster than 'C' version

[vignesh at multicorewareinc.com]

# HG changeset patch
# User Jayashri Murugan <jayashri at multicorewareinc.com>
# Date 1494223284 -19800
#      Mon May 08 11:31:24 2017 +0530
# Node ID 093a7d74fe35ffe5ab4c335efcc690a8e2ed31f2
# Parent  12017e874c9adfd6dbba05a18c3244089d0a51b3
avx2:'integral32v' asm code-> 7.93x faster than 'C' version

    integral_init32v  7.93x    191.18          1516.95

diff -r 12017e874c9a -r 093a7d74fe35 source/common/x86/seaintegral.asm
--- a/source/common/x86/seaintegral.asm	Mon May 08 11:25:18 2017 +0530
+++ b/source/common/x86/seaintegral.asm	Mon May 08 11:31:24 2017 +0530
@@ -133,8 +133,19 @@
 ;void integral_init32v_c(uint32_t *sum32, intptr_t stride)
 ;-----------------------------------------------------------------------------
 INIT_YMM avx2
-cglobal integral32v, 2, 2, 0
- 
+cglobal integral32v, 2, 3, 2
+    mov r2, r1
+    shl r2, 7
+
+.loop
+    movu    m0, [r0]
+    movu    m1, [r0 + r2]
+    psubd   m1, m0
+    movu    [r0], m1
+    add     r0, 32
+    sub     r1, 8
+    cmp     r1, 0
+    jnz     .loop
     RET
 
 ;-----------------------------------------------------------------------------