From 1e7ded69cfaff55c80de1308820031e9a0a01e1a Mon Sep 17 00:00:00 2001
From: "Ronald S. Bultje" <rbultje@google.com>
Date: Wed, 27 Apr 2011 13:49:21 -0700
Subject: [PATCH] Use psadbw to get the sum of bytes in a line.

Thanks Jason for pointing that out on #vp8. ;-).

Change-Id: I5330a753e752a8704b78a409597472628e0b26a5
---
 vp8/common/x86/recon_sse2.asm | 22 ++++------------------
 1 file changed, 4 insertions(+), 18 deletions(-)

diff --git a/vp8/common/x86/recon_sse2.asm b/vp8/common/x86/recon_sse2.asm
index a1cc2b1fef..97dc4f6866 100644
--- a/vp8/common/x86/recon_sse2.asm
+++ b/vp8/common/x86/recon_sse2.asm
@@ -251,15 +251,8 @@ sym(vp8_intra_pred_uv_dc_mmx2):
     movsxd      rax,        dword ptr arg(3) ;src_stride;
     sub         rsi,        rax
     pxor        mm0,        mm0
-    movd        mm1,        [rsi]
-    movd        mm2,        [rsi+4]
-    punpcklbw   mm1,        mm0
-    punpcklbw   mm2,        mm0
-    paddw       mm1,        mm2
-    pshufw      mm2,        mm1, 0x0e
-    paddw       mm1,        mm2
-    pshufw      mm2,        mm1, 0x01
-    paddw       mm1,        mm2
+    movq        mm1,        [rsi]
+    psadbw      mm1,        mm0
 
     ; from left
     dec         rsi
@@ -331,15 +324,8 @@ sym(vp8_intra_pred_uv_dctop_mmx2):
     movsxd      rax,        dword ptr arg(3) ;src_stride;
     sub         rsi,        rax
     pxor        mm0,        mm0
-    movd        mm1,        [rsi]
-    movd        mm2,        [rsi+4]
-    punpcklbw   mm1,        mm0
-    punpcklbw   mm2,        mm0
-    paddw       mm1,        mm2
-    pshufw      mm2,        mm1, 0x0e
-    paddw       mm1,        mm2
-    pshufw      mm2,        mm1, 0x01
-    paddw       mm1,        mm2
+    movq        mm1,        [rsi]
+    psadbw      mm1,        mm0
 
     ; add up
     paddw       mm1,        [GLOBAL(dc_4)]
-- 
GitLab