[x265] [PATCH 2 of 8] asm:intra_pred_ang4_3_sse2 improved ~2.5% 684.95 -> 667.49 with nits and tweaks

[dtyx265 at gmail.com]

# HG changeset patch
# User David T Yuen <dtyx265 at gmail.com>
# Date 1427575220 25200
# Node ID 6595ba5f989fdd521e268911ddf027665a610e25
# Parent  1dfea2291d55ccb628a0f54d7d754812d77213ae
asm:intra_pred_ang4_3_sse2 improved ~2.5% 684.95 -> 667.49 with nits and tweaks

Corrected parameter count
Changed r3 and r4 to r3d and r4d
changed pshufd to psrldq

diff -r 1dfea2291d55 -r 6595ba5f989f source/common/x86/intrapred8.asm
--- a/source/common/x86/intrapred8.asm	Sat Mar 28 13:15:46 2015 -0700
+++ b/source/common/x86/intrapred8.asm	Sat Mar 28 13:40:20 2015 -0700
@@ -1338,11 +1338,11 @@
     RET
 
 INIT_XMM sse2
-cglobal intra_pred_ang4_3, 3,5,8
-    mov         r4, 1
+cglobal intra_pred_ang4_3, 4,5,8
+    mov         r4d, 1
     cmp         r3m, byte 33
-    mov         r3, 9
-    cmove       r3, r4
+    mov         r3d, 9
+    cmove       r3d, r4d
 
     movh        m0, [r2 + r3]   ; [8 7 6 5 4 3 2 1]
     mova        m1, m0
@@ -1404,12 +1404,12 @@
 .store:
     packuswb    m0, m2
     movd        [r0], m0
-    pshufd      m0, m0, 0x39
+    psrldq      m0, 4
     movd        [r0 + r1], m0
-    pshufd      m0, m0, 0x39
+    psrldq      m0, 4
     movd        [r0 + r1 * 2], m0
     lea         r1, [r1 * 3]
-    pshufd      m0, m0, 0x39
+    psrldq      m0, 4
     movd        [r0 + r1], m0
     RET