* GCC Spec2017 optimization Wiki
@ 2019-10-06 14:25 Tamar Christina
2019-10-06 14:53 ` Thomas Koenig
0 siblings, 1 reply; 6+ messages in thread
From: Tamar Christina @ 2019-10-06 14:25 UTC (permalink / raw)
To: gcc; +Cc: nd
Hi All,
As discussed during the Cauldron I have created a wiki page https://gcc.gnu.org/wiki/GCCSpec2017
For us to discuss and exchange ideas to improve GCC's spec score.
I have created pages for all the benchmarks and have filled in an analysis for MCF https://gcc.gnu.org/wiki/GCCSpec2017/mcf
We will be filling in these pages as we go along but I encourage you if you have an analysis or comments to do chime in!
Thanks,
Tamar
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GCC Spec2017 optimization Wiki
2019-10-06 14:25 GCC Spec2017 optimization Wiki Tamar Christina
@ 2019-10-06 14:53 ` Thomas Koenig
2019-10-06 15:00 ` Tamar Christina
0 siblings, 1 reply; 6+ messages in thread
From: Thomas Koenig @ 2019-10-06 14:53 UTC (permalink / raw)
To: Tamar Christina, gcc; +Cc: nd
Am 06.10.19 um 16:25 schrieb Tamar Christina:
> As discussed during the Cauldron I have created a wiki page https://gcc.gnu.org/wiki/GCCSpec2017
> For us to discuss and exchange ideas to improve GCC's spec score.
A few of them are written in Fortran. As SPEC is closed source and
costs are quite high, none of the gfortran maintainers (who are
all volunteers) is in a position to contribute.
Any ideas how to do anything about that?
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: GCC Spec2017 optimization Wiki
2019-10-06 14:53 ` Thomas Koenig
@ 2019-10-06 15:00 ` Tamar Christina
2019-10-06 15:14 ` Thomas Koenig
0 siblings, 1 reply; 6+ messages in thread
From: Tamar Christina @ 2019-10-06 15:00 UTC (permalink / raw)
To: Thomas Koenig, gcc; +Cc: nd
Hi Thomas,
In general our approach is to identify areas for improvement in a benchmark and provide a testcase that's independent of the benchmark when reporting it in a PR upstream.
This simplifies the problem and also allows people who don't have access to SPEC to contribute. Of course this is not always do-able but for a large part of the cases so far this has worked out well.
So I think that's the best way to handle the Fortran and other benchmarks. For the bigger work items we may need to find an alternative.
Cheers,
Tamar
-----Original Message-----
From: Thomas Koenig <tkoenig@netcologne.de>
Sent: Sunday, October 6, 2019 10:53 AM
To: Tamar Christina <Tamar.Christina@arm.com>; gcc@gcc.gnu.org
Cc: nd <nd@arm.com>
Subject: Re: GCC Spec2017 optimization Wiki
Am 06.10.19 um 16:25 schrieb Tamar Christina:
> As discussed during the Cauldron I have created a wiki page
> https://gcc.gnu.org/wiki/GCCSpec2017
> For us to discuss and exchange ideas to improve GCC's spec score.
A few of them are written in Fortran. As SPEC is closed source and costs are quite high, none of the gfortran maintainers (who are all volunteers) is in a position to contribute.
Any ideas how to do anything about that?
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GCC Spec2017 optimization Wiki
2019-10-06 15:00 ` Tamar Christina
@ 2019-10-06 15:14 ` Thomas Koenig
2019-10-06 15:52 ` Toon Moene
0 siblings, 1 reply; 6+ messages in thread
From: Thomas Koenig @ 2019-10-06 15:14 UTC (permalink / raw)
To: Tamar Christina, gcc; +Cc: nd
Hi Tamar,
> In general our approach is to identify areas for improvement in a benchmark and provide a testcase that's independent of the benchmark when reporting it in a PR upstream.
Sounds like a good approach, in principle.
If the people who are doing the identfying know Fortran well, that would
work even better (do they?), and if they could be persuaded to work on
gfortran directly, that would probably be best.
Regards
Thomas
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GCC Spec2017 optimization Wiki
2019-10-06 15:14 ` Thomas Koenig
@ 2019-10-06 15:52 ` Toon Moene
2019-11-07 22:43 ` Toon Moene
0 siblings, 1 reply; 6+ messages in thread
From: Toon Moene @ 2019-10-06 15:52 UTC (permalink / raw)
To: gcc
On 10/6/19 5:14 PM, Thomas Koenig wrote:
> Hi Tamar,
>
>> In general our approach is to identify areas for improvement in a
>> benchmark and provide a testcase that's independent of the benchmark
>> when reporting it in a PR upstream.
>
> Sounds like a good approach, in principle.
>
> If the people who are doing the identfying know Fortran well, that would
> work even better (do they?), and if they could be persuaded to work on
> gfortran directly, that would probably be best.
Of the 7 benchmarks that are (partly) written in Fortran, Cactus is free
software (LGPL'd) and the 3 geological ones (wrf, cam4 and roms) are
"obtainable" (need to register to get the source code). Of course, that
means you get "a" version of the code, not necessarily what is in the
SPEC benchmark, but at least it enables us to join in the analysis.
exchange2 was written by Michael Metcalf, of countless Fortran books,
whom I met once (when I was on the Fortran Standardization Committee).
He might be persuaded to give us a copy for analysis if this really is
an outlier in performance.
Kind regards,
--
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GCC Spec2017 optimization Wiki
2019-10-06 15:52 ` Toon Moene
@ 2019-11-07 22:43 ` Toon Moene
0 siblings, 0 replies; 6+ messages in thread
From: Toon Moene @ 2019-11-07 22:43 UTC (permalink / raw)
To: gcc
Over a month ago, I wrote , about SPEC2017:
> Of the 7 benchmarks that are (partly) written in Fortran, Cactus is free
> software (LGPL'd) and the 3 geological ones (wrf, cam4 and roms) are
> "obtainable" (need to register to get the source code). Of course, that
> means you get "a" version of the code, not necessarily what is in the
> SPEC benchmark, but at least it enables us to join in the analysis.
>
> exchange2 was written by Michael Metcalf, of countless Fortran books,
> whom I met once (when I was on the Fortran Standardization Committee).
> He might be persuaded to give us a copy for analysis if this really is
> an outlier in performance.
Although I still can't completely vouch for it's correctness, I have
written a Sudoku solver (exchange2 is a form of a Sudoku solver) in
Fortran 2018. In principle - if you can find the initial clue
arrangement - it can solve 3x3, 4x4, 5x5, and 6x6 Sudoku's. Up til now,
I have only been able to test it on 3x3 and 4x4 examples.
You'll find it on my web page (indicated below).
--
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2019-11-07 22:43 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-06 14:25 GCC Spec2017 optimization Wiki Tamar Christina
2019-10-06 14:53 ` Thomas Koenig
2019-10-06 15:00 ` Tamar Christina
2019-10-06 15:14 ` Thomas Koenig
2019-10-06 15:52 ` Toon Moene
2019-11-07 22:43 ` Toon Moene
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).