[x265] [PATCH] blockcopy_sp_8x16, optimized asm code

[praveen at multicorewareinc.com]

# HG changeset patch
# User Praveen Tiwari
# Date 1383912504 -19800
# Node ID 227a5666e08869d36e07a75f3db95dd94c774715
# Parent  d9fd009d7118c9cb76b84e94e01732b10d7bf313
blockcopy_sp_8x16, optimized asm code

diff -r d9fd009d7118 -r 227a5666e088 source/common/x86/blockcopy8.asm
--- a/source/common/x86/blockcopy8.asm	Fri Nov 08 17:28:57 2013 +0530
+++ b/source/common/x86/blockcopy8.asm	Fri Nov 08 17:38:24 2013 +0530
@@ -1215,54 +1215,47 @@
 ;-----------------------------------------------------------------------------
 %macro BLOCKCOPY_SP_W8_H8 2
 INIT_XMM sse2
-cglobal blockcopy_sp_%1x%2, 4, 7, 8, dest, destStride, src, srcStride
+cglobal blockcopy_sp_%1x%2, 4, 5, 8, dest, destStride, src, srcStride
 
-mov         r6d,    %2
+mov         r4d,    %2/8
 
 add         r3,     r3
 
-mova        m0,     [tab_Vm]
+.loop
+     movu       m0,      [r2]
+     movu       m1,      [r2 + r3]
+     movu       m2,      [r2 + 2 * r3]
+     lea        r2,      [r2 + 2 * r3]
+     movu       m3,      [r2 + r3]
+     movu       m4,      [r2 + 2 * r3]
+     lea        r2,      [r2 + 2 * r3]
+     movu       m5,      [r2 + r3]
+     movu       m6,      [r2 + 2 * r3]
+     lea        r2,      [r2 + 2 * r3]
+     movu       m7,      [r2 + r3]
 
-.loop
-      movu       m1,     [r2]
-      movu       m2,     [r2 + r3]
-      movu       m3,     [r2 + 2 * r3]
-      lea        r4,     [r2 + 2 * r3]
-      movu       m4,     [r4 + r3]
-      movu       m5,     [r4 + 2 * r3]
-      lea        r4,     [r4 + 2 * r3]
-      movu       m6,     [r4 + r3]
-      movu       m7,     [r4 + 2 * r3]
-      lea        r5,     [r4 + 2 * r3]
+     packuswb   m0,      m1
+     packuswb   m2,      m3
+     packuswb   m4,      m5
+     packuswb   m6,      m7
 
-      pshufb     m1,     m0
-      pshufb     m2,     m0
-      pshufb     m3,     m0
-      pshufb     m4,     m0
-      pshufb     m5,     m0
-      pshufb     m6,     m0
-      pshufb     m7,     m0
+     movlps     [r0],          m0
+     movhps     [r0 + r1],     m0
+     movlps     [r0 + 2 * r1], m2
+     lea        r0,            [r0 + 2 * r1]
+     movhps     [r0 + r1],     m2
+     movlps     [r0 + 2 * r1], m4
+     lea        r0,            [r0 + 2 * r1]
+     movhps     [r0 + r1],     m4
+     movlps     [r0 + 2 * r1], m6
+     lea        r0,            [r0 + 2 * r1]
+     movhps     [r0 + r1],     m6
 
-      movh       [r0],            m1
-      movh       [r0 + r1],       m2
-      movh       [r0 + 2 * r1],   m3
-      lea        r4,              [r0 + 2 * r1]
-      movh       [r4 + r1],       m4
-      movh       [r4 + 2 * r1],   m5
-      lea        r4,              [r4 + 2 * r1]
-      movh       [r4 + r1],       m6
-      movh       [r4 + 2 * r1],   m7
+    lea         r0,            [r0 + 2 * r1]
+    lea         r2,            [r2 + 2 * r3]
 
-      movu       m1,              [r5 + r3]
-      pshufb     m1,              m0
-      lea        r4,              [r4 + 2 * r1]
-      movh       [r4 + r1],       m1
-
-      lea        r0,              [r0 + 8 * r1]
-      lea        r2,              [r2 + 8 * r3]
-
-      sub        r6d,             8
-      jnz        .loop
+    dec         r4d
+    jnz         .loop
 
 RET
 %endmacro