This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] RE: gcc parallel make check
- From: Jakub Jelinek <jakub at redhat dot com>
- To: VandeVondele Joost <joost dot vandevondele at mat dot ethz dot ch>
- Cc: "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>, "fortran at gcc dot gnu dot org" <fortran at gcc dot gnu dot org>, "patches at gcc dot gnu dot org" <patches at gcc dot gnu dot org>
- Date: Fri, 5 Sep 2014 16:53:04 +0200
- Subject: Re: [PATCH] RE: gcc parallel make check
- Authentication-results: sourceware.org; auth=none
- References: <908103EDB4893A42920B21D3568BFD93150F4103 at MBX23 dot d dot ethz dot ch> <20140905143740 dot GL17454 at tucnak dot redhat dot com> <908103EDB4893A42920B21D3568BFD93150F414C at MBX23 dot d dot ethz dot ch>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Fri, Sep 05, 2014 at 02:48:01PM +0000, VandeVondele Joost wrote:
>
> > Please sort the letters (LC_ALL=C sort) and where consecutive, use ranges.
> > Thus \[0-9A-Zhjqvx-z\]* or so. What is - doing in your list? _-9 is quite
> > unexpected range.
>
> the '-' is a bug indeed. I added this to the list of allowed starting characters, but this clearly this is not a good idea. Thanks, removed.
>
> > Also, please tweak the script so that it will do the
> > sorting and ranges handling for you.
>
> The current order is the number of files that start with this letter, which is somewhat informative, but let me see if I can have ranges easily (in awk...).
BTW, number of tests might help as initial guess on how to split things, but
might be severely off.
What I'd recommend after you split with your script that you perform a
make -j1 check and diff the Test run by and runtest completed
times in each of the */*.log.sep files and see what time each job takes,
then perhaps reshuffle stuff a little bit until it is roughly even.
Some letters could start more dg-do run tests while others more dg-do
compile once, some tests take a short time to compile, some very long, etc.
Jakub