The /z/m5 directory is on an internal U-M machine (zizzer), though the
regressions themselves are typically run on the Michigan simulation pool.
For this particular case, that shouldn't matter though; you should get the
same failure when you run the same test on a clean repository.  If not, then
the problem is probably in the regression environment and not in the
simulator.

Steve

On Sun, Jul 12, 2009 at 5:12 PM, Derek Hower <[email protected]> wrote:

> Where is the regression run?  I logged on repo.m5sim.org to look at
> the log file, but there is no /z/m5 directory.
>
> -Derek
>
> On Sat, Jul 11, 2009 at 10:11 AM, nathan binkert <[email protected]> wrote:
> >
> > > *****
> build/ALPHA_SE/tests/fast/quick/50.memtest/alpha/linux/memtest-ruby FAILED!
> >
> > Derek,
> >
> >
> > M5 regressions not only ensure that the program doesn't crash, but
> > they ensure that the stats don't change either.  So, this stat likely
> > failed because the stats changed somewhat due to your recent fix.
> > Basically you need to run the test look at the changes if they're
> > right, re-run the test with update_ref=True added to the SCons command
> > line (it shouldn't actually run again if nothing changed).  This will
> > copy new reference files to the tests directory and you should then
> > commit them.
> >
> > Make sense?
> >
> >  Nate
> > _______________________________________________
> > m5-dev mailing list
> > [email protected]
> > http://m5sim.org/mailman/listinfo/m5-dev
> _______________________________________________
> m5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/m5-dev
>
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to