On Mon, Mar 2, 2015 at 4:37 PM, Bill Schmidt
wrote:
> Hi,
>
> On Thu, 2015-02-26 at 16:48 -0600, Bill Schmidt wrote:
>>
>> Except this didn't pass regression testing. I'll continue to poke at
>> it.
>
> I figured out the issue here. My handling of the
> TARGET_ALLOW_MOVMISALIGN and TARGET_EFFICI
Hi,
On Thu, 2015-02-26 at 16:48 -0600, Bill Schmidt wrote:
>
> Except this didn't pass regression testing. I'll continue to poke at
> it.
I figured out the issue here. My handling of the
TARGET_ALLOW_MOVMISALIGN and TARGET_EFFICIENT_UNALIGNED_VSX macros was
correct, but there was another probl
On Thu, 2015-02-26 at 14:40 -0600, Bill Schmidt wrote:
> On Thu, 2015-02-26 at 10:36 -0500, David Edelsohn wrote:
>
> > My one concern is the interaction between TARGET_ALLOW_MOVMISALIGN and
> > TARGET_EFFICIENT_UNALIGNED_VSX in the movmisalign pattern in
> > vector.md. Your patch changes
> > rs6
On Thu, 2015-02-26 at 10:36 -0500, David Edelsohn wrote:
> My one concern is the interaction between TARGET_ALLOW_MOVMISALIGN and
> TARGET_EFFICIENT_UNALIGNED_VSX in the movmisalign pattern in
> vector.md. Your patch changes
> rs6000_builtin_support_vector_misalignment to return TRUE if
> TARGET_
On Tue, Feb 10, 2015 at 8:18 PM, Bill Schmidt
wrote:
> Hi David,
>
> Sorry I haven't gotten back to this sooner. I've tried to address your
> comments about the previous patch. Please let me know if I'm off base.
> Test results are the same as previously.
>
> Thanks,
> Bill
>
>
> [gcc]
>
> 2015-
Hi David,
Sorry I haven't gotten back to this sooner. I've tried to address your
comments about the previous patch. Please let me know if I'm off base.
Test results are the same as previously.
Thanks,
Bill
[gcc]
2015-02-10 Bill Schmidt
* config/rs6000/rs6000.c (rs6000_option_over
On Jan 29, 2015, at 11:39 AM, David Edelsohn wrote:
>
> This is a reasonable change, but please ask a vect maintainer like
> Richi or a testsuite maintainer to approve.
So, my perspective, looks usual and customary.
I do worry about the turing complete nature of testing and the endless lets
bu
On Fri, Jan 23, 2015 at 12:23 PM, Bill Schmidt
wrote:
> Hi,
>
> The POWER8 processor greatly improves performance of unaligned vector
> loads and stores. Except for certain corner cases the compiler can't
> readily track, an unaligned vector load or store performs equivalently
> to an aligned one
Hi,
The POWER8 processor greatly improves performance of unaligned vector
loads and stores. Except for certain corner cases the compiler can't
readily track, an unaligned vector load or store performs equivalently
to an aligned one.
To exploit this in the auto-vectorizer requires two changes. T
