[x265] [PATCH] weighted prediction (pixel), interface simplification

[chen]

At 2014-10-17 18:29:49,praveen at multicorewareinc.com wrote:
># HG changeset patch
># User Praveen Tiwari
># Date 1413541750 -19800
># Node ID 61051f5a16b387120b17be2024543d14aea61f16
># Parent  b7eeae24aae63495bcad1570ecd76cae988f0f6e
>weighted prediction (pixel), interface simplification
>
> void x265_pixel_ssim_4x4x2_core_mmx2(const uint8_t * pix1, intptr_t stride1,
>diff -r b7eeae24aae6 -r 61051f5a16b3 source/common/x86/pixel-util8.asm
>--- a/source/common/x86/pixel-util8.asm	Thu Oct 16 21:57:30 2014 +0530
>+++ b/source/common/x86/pixel-util8.asm	Fri Oct 17 15:59:10 2014 +0530
>@@ -1298,35 +1298,32 @@
> 
> 
> ;-----------------------------------------------------------------------------------------------------------------------------------------------
>-;void weight_pp(pixel *src, pixel *dst, intptr_t srcStride, intptr_t dstStride, int width, int height, int w0, int round, int shift, int offset)
>+;void weight_pp(pixel *src, pixel *dst, intptr_t stride, int width, int height, int w0, int round, int shift, int offset)
> ;-----------------------------------------------------------------------------------------------------------------------------------------------
> INIT_XMM sse4
> cglobal weight_pp, 6, 7, 6
> 
>+    shl         r5d, 6      ; m0 = [w0<<6]
>     mov         r6d, r6m
>-    shl         r6d, 6
>-    movd        m0, r6d         ; m0 = [w0<<6]
>-
>-    movd        m1, r7m         ; m1 = [round]
>-    punpcklwd   m0, m1          ; assuming both (w0<<6) and round are using maximum of 16 bits each.
>-    pshufd      m0, m0, 0       ; m0 = [w0<<6 round]
>-
>-    movd        m1, r8m
>-
>-    movd        m2, r9m
>+    shl         r6d, 16
>+    or          r6d, r5d    ; assuming both (w0<<6) and round are using maximum of 16 bits each.

Please modify C model with assert()
 
>+    movd        m0, r6d
>+    pshufd      m0, m0, 0   ; m0 = [w0<<6, round]
>+
>+    movd        m1, r7m
>+
>+    movd        m2, r8m
>     pshufd      m2, m2, 0
> 
>     mova        m5, [pw_1]
> 
>-    sub         r2d, r4d
>-    sub         r3d, r4d
>-
>+    sub         r2d, r3d
>+    shr         r3d, 4
> .loopH:
>-    mov         r6d, r4d
>-    shr         r6d, 4
>+    mov         r5d, r3d
>+
> .loopW:
>-    movh        m4, [r0]
>-    pmovzxbw    m4, m4
>+    pmovzxbw    m4, [r0]
> 
>     punpcklwd   m3, m4, m5
>     pmaddwd     m3, m0
>@@ -1364,13 +1361,13 @@
>     add         r0, 16
>     add         r1, 16
> 
>-    dec         r6d
>+    dec         r5d
>     jnz         .loopW
> 
>     lea         r0, [r0 + r2]
>-    lea         r1, [r1 + r3]
>-
>-    dec         r5d
>+    lea         r1, [r1 + r2]
>+
>+    dec         r4d
>     jnz         .loopH
> 
>     RET
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20141018/97f499d0/attachment.html>