On Thu, Feb 11, 2016 at 5:30 PM, chen <[email protected]> wrote:
>
> At 2016-02-11 17:54:45,[email protected] wrote:
> ># HG changeset patch
> ># User [email protected]
> ># Date 1455183020 -19800
> ># Thu Feb 11 15:00:20 2016 +0530
> ># Node ID 4f5720ccaf1aa04868054636f14dce8ea65390ad
> ># Parent a2ff6747eaf7b25102f27f808cf5526f441df488
> >arm: Implement blockcopy_pp_NxN_neon
> >
> >+function x265_blockcopy_pp_48x64_neon
> >+ push {r4, r5}
> >+ mov r4, #8
> >+ mov r5, #32
> >+ sub r3, r5
> >+ sub r1, r5
> >+loop_48x64:
> >+.rept 8
> >+ vld1.8 {q0, q1}, [r2]!
> >+ vld1.8 {q2}, [r2], r3
> the ARM support format "vld1.8 {q0, q1, q1}, Rn, Rm"
>
>
> Load support maximum upto 4 double word vld1.8 {d0,d1,d2,d3}, [Rn], Rm
> OR 2 quadwords vld1.8 {q0,q1}, [Rn], Rm
>
>
> >+ vst1.8 {q0, q1}, [r0]!
> >+ vst1.8 {q2}, [r0], r1
> >+.endr
> >+ subs r4, r4, #1
> >+ bne loop_48x64
> >+ pop {r4, r5}
> >+ bx lr
> >+endfunc
>
>
> _______________________________________________
> x265-devel mailing list
> [email protected]
> https://mailman.videolan.org/listinfo/x265-devel
>
>
_______________________________________________
x265-devel mailing list
[email protected]
https://mailman.videolan.org/listinfo/x265-devel
