Hi Siarhei,
It is always possible to find time for fixing bugs and reviewing the
code, which is doing something obviously useful.
I think we have been unlucky because it seems to me that our schedules
haven't aligned very well. I have put time aside on a few occasions over
the last couple of y
On Thu, 7 Apr 2016 19:45:03 +0900
Mizuki Asakura wrote:
> > Do you have a more specific example of a code fragment that needs
> > conversion?
>
> In original pixman-arm-neon-asm.S:
>
> .macro pixman_composite_over__8_0565_process_pixblock_head
> ...
> vsli.u16q2, q2, #5
> ...
> vradd
> Do you have a more specific example of a code fragment that needs
> conversion?
In original pixman-arm-neon-asm.S:
.macro pixman_composite_over__8_0565_process_pixblock_head
...
vsli.u16q2, q2, #5
...
vraddhn.u16 d2, q6, q10
...
vshrn.u16 d30, q2, #2
If all registers just convert
On Tue, 5 Apr 2016 19:28:19 +0900
Mizuki Asakura wrote:
> > It looks like you have used an automated process to convert the AArch32
> > NEON code to AArch64. Will you be able to repeat that process for other
> > code, or at least assist others to repeat your steps?
>
> Sorry, but I've wrote be
On Tue, 5 Apr 2016 20:20:54 +0900
Mizuki Asakura wrote:
> > This code is not just there for prefetching. It is an example of
> > using software pipelining:
>
> OK. I understand.
> But the code is very hard to maintain... I've met too many register
> conflictions.
The *_tail_head variant has e