Move sub before store to avoid wait on core2.

BUG=101
TEST=none
Review URL: https://webrtc-codereview.appspot.com/846006

git-svn-id: http://libyuv.googlecode.com/svn/trunk@377 16f28f9a-4ce2-e073-06de-1de4eb20be90

README.chromium

 Name: libyuv
 URL: http://code.google.com/p/libyuv/
-Version: 375
+Version: 377
 License: BSD
 License File: LICENSE
 

include/libyuv/version.h

@@ -11,6 +11,6 @@
 #ifndef INCLUDE_LIBYUV_VERSION_H_  // NOLINT
 #define INCLUDE_LIBYUV_VERSION_H_
 
-#define LIBYUV_VERSION 375
+#define LIBYUV_VERSION 377
 
 #endif  // INCLUDE_LIBYUV_VERSION_H_  NOLINT

source/row_win.cc

@@ -679,9 +679,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqa     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }
@@ -713,9 +713,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqu     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }
@@ -747,9 +747,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqa     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }
@@ -781,9 +781,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqu     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }
@@ -815,9 +815,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqa     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }
@@ -883,9 +883,9 @@ __asm {
     psrlw      xmm2, 7
     packuswb   xmm0, xmm2
     paddb      xmm0, xmm5
+    sub        ecx, 16
     movdqa     [edx], xmm0
     lea        edx, [edx + 16]
-    sub        ecx, 16
     jg         convertloop
     ret
   }