ffc team mailing list archive
-
ffc team
-
Mailing list archive
-
Message #02183
Re: buildbot failure in FEniCS Buildbot on ffc-linux-64
Quoting Anders Logg <logg@xxxxxxxxx>:
> On Wed, Mar 04, 2009 at 12:39:31PM +0100, Kristian Oelgaard wrote:
> > Quoting buildbot@xxxxxxxxxx:
> >
> > > The Buildbot has detected a new failure of ffc-linux-64 on FEniCS
> Buildbot.
> > > Full details are available at:
> > > http://fenics.org:8080/builders/ffc-linux-64/builds/3
> > >
> > > Buildbot URL: http://fenics.org:8080/
> > >
> > > Buildslave for this Build: linux-64
> > >
> > > Build Reason:
> > > Build Source Stamp: HEAD
> > > Blamelist: k.b.oelgaard
> > >
> > > BUILD FAILED: failed ffc check
> >
> > I don't get this error. It complains about the JIT compiler not being fast
> > enough, but it is OK on my computer.
> >
> > JIT in-memory cache: 9.08374786377e-05
> > JIT disk cache: 0.116662979126
> > Reasonable values are 0.002 and 0.03
>
> Something seems to have happened with the disk cache. You get 0.1
> seconds. I have
>
> JIT in-memory cache: 9.70363616943e-05
> JIT disk cache: 0.0144801139832
Yes, I get similar numbers on my machine. The numbers I reported above is from
the buildbot.
> The disk cache involves computing the signature. Has something change
> wrt computing signatures?
Not as far as I know.
Kristian
> --
> Anders
>
> > F..
> > ======================================================================
> > FAIL: Test that JIT compiler is fast enough.
> > ----------------------------------------------------------------------
> > Traceback (most recent call last):
> > File "test.py", line 254, in testPoisson
> > self.assertTrue(dt1 < 0.1) ## Is this threshold too strict??
> > AssertionError
> >
> > Kristian
> >
> > > sincerely,
> > > -The Buildbot
> > >
> > > _______________________________________________
> > > FFC-dev mailing list
> > > FFC-dev@xxxxxxxxxx
> > > http://www.fenics.org/mailman/listinfo/ffc-dev
> > >
> >
> >
> > _______________________________________________
> > FFC-dev mailing list
> > FFC-dev@xxxxxxxxxx
> > http://www.fenics.org/mailman/listinfo/ffc-dev
>
Follow ups
References