Faster integral_init

palignr to avoid unaligned loads is worth it in inith, but not initv.

diff --git a/common/x86/mc-a2.asm b/common/x86/mc-a2.asm
index 89e1212a934ffb507eca08d63f167c0060be93bb..cebe59fdc0f28c566dfe7f9cd47a65e38ed9b51c 100644 (file)
--- a/common/x86/mc-a2.asm
+++ b/common/x86/mc-a2.asm
@@ -759,7 +759,8 @@ cglobal x264_integral_init4h_sse4, 3,4
     pxor    m4, m4
 .loop:
     movdqa  m0, [r1+r2]
-    movdqu  m1, [r1+r2+8]
+    movdqa  m1, [r1+r2+16]
+    palignr m1, m0, 8
     mpsadbw m0, m4, 0
     mpsadbw m1, m4, 0
     paddw   m0, [r0+r2*2]
@@ -777,7 +778,8 @@ cglobal x264_integral_init8h_sse4, 3,4
     pxor    m4, m4
 .loop:
     movdqa  m0, [r1+r2]
-    movdqu  m1, [r1+r2+8]
+    movdqa  m1, [r1+r2+16]
+    palignr m1, m0, 8
     movdqa  m2, m0
     movdqa  m3, m1
     mpsadbw m0, m4, 0