On 6/10/25 1:55 PM, Harald Anlauf wrote:
Gesendet: Mittwoch, 4. Juni 2025 um 20:37
Von: "Jerry D" <jvdelis...@gmail.com>
An: "Mikael Morin" <morin-mik...@orange.fr>, "Harald Anlauf" <anl...@gmx.de>,
fortran@gcc.gnu.org
Betreff: Re: Execution time for gfortran regression testing
On 6/3/25 3:02 AM, Mikael Morin wrote:
The better direction is probably to make that TORTURE_OPTIONS trick
work, so that you can get results fast when you ask for it, but the
overall coverage of options testing is not reduced in general.
I've just spent some time to look at this TORTURE_OPTIONS thing. Here is
how I think it works.
1. create the file site.exp in the build directory:
$ make -C gcc site.exp
2. open the just-created gcc/site.exp file add this line at the bottom:
set TORTURE_OPTIONS [list { -O2 }]
3. save the file
4. run the testsuite
According to my (limited) testing this does reduce the number of
executions to a single one.
I hope it helps.
This is really a great suggestion! It really helps.
I tried this here and it does speed things up. I am running:
make -k -j10 check-fortran
Before the change:
Testing of trunk complete.....
real 4m57.587s
user 36m32.709s
sys 12m33.567s
After the change:
Testing of trunk complete.....
real 1m14.890s
user 8m56.192s
sys 3m23.847s
That is a significant speedup.
My machine is about an order of magnitude slower...
(Nowadays I need of the order of 1.5 hours for full testing of trunk).
Harald
How about a script to scan through the testsuite and do a compile test
only? I could do this on the { dg-do compile } or { dg-do run } files at
-O2 only. I don't mind taking the time if it helps everyone else be more
efficient. ( yes, I am a glutton for punishment ;) )
Jerry
