Hi Daniel!
On Wed, Nov 18, 2009 at 03:32:25PM -0800, William Stein wrote:
> On Wed, Nov 18, 2009 at 3:29 PM, Daniel Bump <b...@match.stanford.edu> wrote:
> > The patch #5794 adds a lot of branching rules, and tests are provided
> > for most of these. I marked many of them long.
> >
> > I won't have time immediately but should be able to get to it
> > over the weekend. If that is not soon enough, I don't mind if
> > you do it. In case you do, observe the following lines:
> >
> > sage: E6=WeylCharacterRing("E6",style="coroots") # long time
> > sage: D5=WeylCharacterRing("D5",style="coroots") # long time
> > sage: fw = E6.fundamental_weights() # long time
> > sage: [E6(fw[i]).branch(D5,rule="levi") for i in [1,2,6]] # long time
> >
> > I think only the last item in this list will be slow. (Actually
> > this test takes about 1.99 seconds on one machine.) But unless they
> > are involved in some test, there is no need to define the Weyl
> > character rings D5 and E6, so the other lines are marked long time too.
> >
> > As I read the doc, no single test should exceed 30s. But there are
> > tests marked "long time". So it is possible that most of them
> > are under 30s yet they accumulate to over 250s. Therefore could
> > you suggest a target for the entire file to run on with typical
> > hardware when testing with -long?
I just ran each long test individually, and annotated them with how
much time they took (on my macbook pro). They are indeed all below
30s, except for a big E8 test which takes 160s. I marked that one as
#not tested, and the file now runs under 150s which is good enough
(especially since I expect further serious improvements when we will
optimize the combinatorial free module code).
I removed the #long time on a couple calculations that were below
0.7s. I also removed the #long time on the declarations (like the
E6=...). They don't cost much, and one never knows; they might catch
something.
Dan or Mike: please review trac_5794-long-time-nt.patch on #5794!
By the way: I posted a feature request on trac to streamline this kind
of doctest timing analysis: #7493
Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nthi...@users.sf.net>
http://Nicolas.Thiery.name/
--
You received this message because you are subscribed to the Google Groups
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-de...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/sage-combinat-devel?hl=.
