[x265] [PATCH] asm: 16bpp code for quant and dequant_normal

[Murugan Vairavel]

yes, scale value less than 32768 should take the faster path and scale
value >= 32768 should take the slower path. I will send patch for that.



On Thu, Feb 20, 2014 at 12:02 AM, chen <chenm003 at 163.com> wrote:

> >@@ -1154,8 +1151,11 @@
>
> > ; void dequant_normal(const int32_t* quantCoef, int32_t* coef, int num, int scale, int shift)
>
> > ;-----------------------------------------------------------------------------
> > INIT_XMM sse4
> >-cglobal dequant_normal, 2,5,8
> >-    movd        m1, r3m             ; m1 = word [scale]
> >+cglobal dequant_normal, 4,5,5
> >+    movd        m1, r3             ; m1 = word [scale]
> >+    cmp         r3d, 255
> r3 is scale, it maximum is (72 << per), when QP more than 12, the code
> will enter slower path.
>
>
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
>


-- 
With Regards,

Murugan. V
+919659287478
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20140220/a9d07da3/attachment.html>