* [PATCH] Change fma_reassoc_width tuning for ampere1
@ 2023-06-20 3:58 Di Zhao OS
2023-06-22 14:17 ` Richard Sandiford
0 siblings, 1 reply; 5+ messages in thread
From: Di Zhao OS @ 2023-06-20 3:58 UTC (permalink / raw)
To: gcc-patches
This patch enables reassociation of floating-point additions on ampere1.
This brings about 1% overall benefit on spec2017 fprate cases. (There
are minor regressions in 510.parest_r and 508.namd_r, analyzed here:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110279 .)
Bootstrapped and tested on aarch64-unknown-linux-gnu. Is this OK for trunk?
Thanks,
Di Zhao
gcc/ChangeLog:
* config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1
---
diff --git a/gcc/config/aarch64/aarch64.cc b/gcc/config/aarch64/aarch64.cc
index d16565b5581..301c9f6c0cd 100644
--- a/gcc/config/aarch64/aarch64.cc
+++ b/gcc/config/aarch64/aarch64.cc
@@ -1927,7 +1927,7 @@ static const struct tune_params ampere1_tunings =
"32:12", /* loop_align. */
2, /* int_reassoc_width. */
4, /* fp_reassoc_width. */
- 1, /* fma_reassoc_width. */
+ 4, /* fma_reassoc_width. */
2, /* vec_reassoc_width. */
2, /* min_div_recip_mul_sf. */
2, /* min_div_recip_mul_df. */
--
2.25.1
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] Change fma_reassoc_width tuning for ampere1
2023-06-20 3:58 [PATCH] Change fma_reassoc_width tuning for ampere1 Di Zhao OS
@ 2023-06-22 14:17 ` Richard Sandiford
2023-06-22 14:23 ` Philipp Tomsich
0 siblings, 1 reply; 5+ messages in thread
From: Richard Sandiford @ 2023-06-22 14:17 UTC (permalink / raw)
To: Di Zhao OS via Gcc-patches; +Cc: Di Zhao OS
Di Zhao OS via Gcc-patches <gcc-patches@gcc.gnu.org> writes:
> This patch enables reassociation of floating-point additions on ampere1.
> This brings about 1% overall benefit on spec2017 fprate cases. (There
> are minor regressions in 510.parest_r and 508.namd_r, analyzed here:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110279 .)
>
> Bootstrapped and tested on aarch64-unknown-linux-gnu. Is this OK for trunk?
>
> Thanks,
> Di Zhao
>
> gcc/ChangeLog:
>
> * config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1
Thanks, pushed to trunk.
Richard
> ---
> diff --git a/gcc/config/aarch64/aarch64.cc b/gcc/config/aarch64/aarch64.cc
> index d16565b5581..301c9f6c0cd 100644
> --- a/gcc/config/aarch64/aarch64.cc
> +++ b/gcc/config/aarch64/aarch64.cc
> @@ -1927,7 +1927,7 @@ static const struct tune_params ampere1_tunings =
> "32:12", /* loop_align. */
> 2, /* int_reassoc_width. */
> 4, /* fp_reassoc_width. */
> - 1, /* fma_reassoc_width. */
> + 4, /* fma_reassoc_width. */
> 2, /* vec_reassoc_width. */
> 2, /* min_div_recip_mul_sf. */
> 2, /* min_div_recip_mul_df. */
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] Change fma_reassoc_width tuning for ampere1
2023-06-22 14:17 ` Richard Sandiford
@ 2023-06-22 14:23 ` Philipp Tomsich
2023-06-26 14:27 ` Richard Sandiford
0 siblings, 1 reply; 5+ messages in thread
From: Philipp Tomsich @ 2023-06-22 14:23 UTC (permalink / raw)
To: Richard Sandiford, Di Zhao OS via Gcc-patches, Di Zhao OS
Richard,
OK for backport to GCC-13?
Thanks,
Philipp.
On Thu, 22 Jun 2023 at 16:18, Richard Sandiford via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> Di Zhao OS via Gcc-patches <gcc-patches@gcc.gnu.org> writes:
> > This patch enables reassociation of floating-point additions on ampere1.
> > This brings about 1% overall benefit on spec2017 fprate cases. (There
> > are minor regressions in 510.parest_r and 508.namd_r, analyzed here:
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110279 .)
> >
> > Bootstrapped and tested on aarch64-unknown-linux-gnu. Is this OK for trunk?
> >
> > Thanks,
> > Di Zhao
> >
> > gcc/ChangeLog:
> >
> > * config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1
>
> Thanks, pushed to trunk.
>
> Richard
>
> > ---
> > diff --git a/gcc/config/aarch64/aarch64.cc b/gcc/config/aarch64/aarch64.cc
> > index d16565b5581..301c9f6c0cd 100644
> > --- a/gcc/config/aarch64/aarch64.cc
> > +++ b/gcc/config/aarch64/aarch64.cc
> > @@ -1927,7 +1927,7 @@ static const struct tune_params ampere1_tunings =
> > "32:12", /* loop_align. */
> > 2, /* int_reassoc_width. */
> > 4, /* fp_reassoc_width. */
> > - 1, /* fma_reassoc_width. */
> > + 4, /* fma_reassoc_width. */
> > 2, /* vec_reassoc_width. */
> > 2, /* min_div_recip_mul_sf. */
> > 2, /* min_div_recip_mul_df. */
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] Change fma_reassoc_width tuning for ampere1
2023-06-22 14:23 ` Philipp Tomsich
@ 2023-06-26 14:27 ` Richard Sandiford
2023-07-29 14:22 ` Di Zhao OS
0 siblings, 1 reply; 5+ messages in thread
From: Richard Sandiford @ 2023-06-26 14:27 UTC (permalink / raw)
To: Philipp Tomsich; +Cc: Di Zhao OS via Gcc-patches, Di Zhao OS
Philipp Tomsich <philipp.tomsich@vrull.eu> writes:
> Richard,
>
> OK for backport to GCC-13?
Yeah, OK for GCC 13 too.
Thanks,
Richard
> Thanks,
> Philipp.
>
> On Thu, 22 Jun 2023 at 16:18, Richard Sandiford via Gcc-patches
> <gcc-patches@gcc.gnu.org> wrote:
>>
>> Di Zhao OS via Gcc-patches <gcc-patches@gcc.gnu.org> writes:
>> > This patch enables reassociation of floating-point additions on ampere1.
>> > This brings about 1% overall benefit on spec2017 fprate cases. (There
>> > are minor regressions in 510.parest_r and 508.namd_r, analyzed here:
>> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110279 .)
>> >
>> > Bootstrapped and tested on aarch64-unknown-linux-gnu. Is this OK for trunk?
>> >
>> > Thanks,
>> > Di Zhao
>> >
>> > gcc/ChangeLog:
>> >
>> > * config/aarch64/aarch64.cc: Change fma_reassoc_width for ampere1
>>
>> Thanks, pushed to trunk.
>>
>> Richard
>>
>> > ---
>> > diff --git a/gcc/config/aarch64/aarch64.cc b/gcc/config/aarch64/aarch64.cc
>> > index d16565b5581..301c9f6c0cd 100644
>> > --- a/gcc/config/aarch64/aarch64.cc
>> > +++ b/gcc/config/aarch64/aarch64.cc
>> > @@ -1927,7 +1927,7 @@ static const struct tune_params ampere1_tunings =
>> > "32:12", /* loop_align. */
>> > 2, /* int_reassoc_width. */
>> > 4, /* fp_reassoc_width. */
>> > - 1, /* fma_reassoc_width. */
>> > + 4, /* fma_reassoc_width. */
>> > 2, /* vec_reassoc_width. */
>> > 2, /* min_div_recip_mul_sf. */
>> > 2, /* min_div_recip_mul_df. */
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: [PATCH] Change fma_reassoc_width tuning for ampere1
2023-06-26 14:27 ` Richard Sandiford
@ 2023-07-29 14:22 ` Di Zhao OS
0 siblings, 0 replies; 5+ messages in thread
From: Di Zhao OS @ 2023-07-29 14:22 UTC (permalink / raw)
To: Richard Sandiford, Philipp Tomsich; +Cc: gcc-patches
Cherry-picked this to gcc-13.
Thanks,
Di Zhao
> -----Original Message-----
> From: Richard Sandiford <richard.sandiford@arm.com>
> Sent: Monday, June 26, 2023 10:28 PM
> To: Philipp Tomsich <philipp.tomsich@vrull.eu>
> Cc: Di Zhao OS via Gcc-patches <gcc-patches@gcc.gnu.org>; Di Zhao OS
> <dizhao@os.amperecomputing.com>
> Subject: Re: [PATCH] Change fma_reassoc_width tuning for ampere1
>
> Philipp Tomsich <philipp.tomsich@vrull.eu> writes:
> > Richard,
> >
> > OK for backport to GCC-13?
>
> Yeah, OK for GCC 13 too.
>
> Thanks,
> Richard
>
> > Thanks,
> > Philipp.
> >
> > On Thu, 22 Jun 2023 at 16:18, Richard Sandiford via Gcc-patches
> > <gcc-patches@gcc.gnu.org> wrote:
> >>
> >> Di Zhao OS via Gcc-patches <gcc-patches@gcc.gnu.org> writes:
> >> > This patch enables reassociation of floating-point additions on ampere1.
> >> > This brings about 1% overall benefit on spec2017 fprate cases. (There
> >> > are minor regressions in 510.parest_r and 508.namd_r, analyzed here:
> >> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110279 .)
> >> >
> >> > Bootstrapped and tested on aarch64-unknown-linux-gnu. Is this OK for
> trunk?
> >> >
> >> > Thanks,
> >> > Di Zhao
> >> >
> >> > gcc/ChangeLog:
> >> >
> >> > * config/aarch64/aarch64.cc: Change fma_reassoc_width for
> ampere1
> >>
> >> Thanks, pushed to trunk.
> >>
> >> Richard
> >>
> >> > ---
> >> > diff --git a/gcc/config/aarch64/aarch64.cc
> b/gcc/config/aarch64/aarch64.cc
> >> > index d16565b5581..301c9f6c0cd 100644
> >> > --- a/gcc/config/aarch64/aarch64.cc
> >> > +++ b/gcc/config/aarch64/aarch64.cc
> >> > @@ -1927,7 +1927,7 @@ static const struct tune_params ampere1_tunings =
> >> > "32:12", /* loop_align. */
> >> > 2, /* int_reassoc_width. */
> >> > 4, /* fp_reassoc_width. */
> >> > - 1, /* fma_reassoc_width. */
> >> > + 4, /* fma_reassoc_width. */
> >> > 2, /* vec_reassoc_width. */
> >> > 2, /* min_div_recip_mul_sf. */
> >> > 2, /* min_div_recip_mul_df. */
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-07-29 14:22 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-06-20 3:58 [PATCH] Change fma_reassoc_width tuning for ampere1 Di Zhao OS
2023-06-22 14:17 ` Richard Sandiford
2023-06-22 14:23 ` Philipp Tomsich
2023-06-26 14:27 ` Richard Sandiford
2023-07-29 14:22 ` Di Zhao OS
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).