[Bug fortran/82362] [8 Regression] SPEC CPU2006 436.cactusADM ~7% performance deviation with trunk@251713

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82362

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |x86_64-*-*, i?86-*-*
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|2017-09-30 00:00:00         |2017-10-02
          Component|middle-end                  |fortran
     Ever confirmed|0                           |1
           Severity|normal                      |enhancement

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I think the Fortran FE could help with using alloca-with-align with larger
alignment (maybe using BIGGEST_ALIGNMENT, I see really no good reason for
using sth as large as 64 bytes by default, so yes, for this we'd need an
option that amends -fstack-arrays).

But that Fortran part isn't really a regression.

> Comparing performance for particular 7-runs
> (relative to median performance of r251711):
> r251711: 92,8%   92,9%   93,0%   106,7%  107,0%  107,0%  107,2%
> r251713: 99,5%   99,6%   99,8%   100,0%  100,3%  100,6%  100,6%

I suppose you swapped the revs here as here r251713 looks more stable.

The median looks unchanged here so the regression would be the instability.