<div dir="ltr"><br><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Oct 28, 2013 at 5:46 AM,  <span dir="ltr"><<a href="mailto:dnyaneshwar@multicorewareinc.com" target="_blank">dnyaneshwar@multicorewareinc.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"># HG changeset patch<br>
# User Dnyaneshwar Gorade <<a href="mailto:dnyaneshwar@multicorewareinc.com">dnyaneshwar@multicorewareinc.com</a>><br>
# Date 1382957168 -19800<br>
#      Mon Oct 28 16:16:08 2013 +0530<br>
# Node ID a44e10ccd754dcd8a5c24bc078c1912fa80782a0<br>
# Parent  b9a9f725d4892ced1176288fe1e20069fa188132<br>
asm: assembly code for pixel_sad_16x64<br></blockquote><div><br></div><div>this also causes encoder crashes</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

diff -r b9a9f725d489 -r a44e10ccd754 source/common/x86/asm-primitives.cpp<br>
--- a/source/common/x86/asm-primitives.cpp      Mon Oct 28 16:11:20 2013 +0530<br>
+++ b/source/common/x86/asm-primitives.cpp      Mon Oct 28 16:16:08 2013 +0530<br>
@@ -236,6 +236,7 @@<br>
         //PIXEL_AVE(sse2);<br>
<br>
         p.sad[LUMA_16x32]  = x265_pixel_sad_16x32_sse2;<br>
+        p.sad[LUMA_16x64]  = x265_pixel_sad_16x64_sse2;<br>
<br>
         ASSGN_SSE(sse2);<br>
         INIT2(sad, _sse2);<br>
diff -r b9a9f725d489 -r a44e10ccd754 source/common/x86/sad-a.asm<br>
--- a/source/common/x86/sad-a.asm       Mon Oct 28 16:11:20 2013 +0530<br>
+++ b/source/common/x86/sad-a.asm       Mon Oct 28 16:16:08 2013 +0530<br>
@@ -267,6 +267,34 @@<br>
     movd   eax, m0<br>
     RET<br>
<br>
+;-----------------------------------------------------------------------------<br>
+; int pixel_sad_16x64( uint8_t *, intptr_t, uint8_t *, intptr_t )<br>
+;-----------------------------------------------------------------------------<br>
+cglobal pixel_sad_16x64, 4,4,3<br>
+    pxor m0, m0<br>
+<br>
+    PROCESS_SAD_4x16<br></blockquote><div><br></div><div>4x16 seems like an innefficient way to split up 16x64</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+    PROCESS_SAD_4x16<br>
+<br>
+    movhlps m1, m0<br>
+    paddd   m0, m1<br>
+    movd   eax, m0<br>
+    RET<br>
+<br>
 %endmacro<br>
<br>
 INIT_XMM sse2<br>
_______________________________________________<br>
x265-devel mailing list<br>
<a href="mailto:x265-devel@videolan.org">x265-devel@videolan.org</a><br>
<a href="https://mailman.videolan.org/listinfo/x265-devel" target="_blank">https://mailman.videolan.org/listinfo/x265-devel</a><br>
</blockquote></div><br><br clear="all"><div><br></div>-- <br>Steve Borho
</div></div>