On Dec 4, 4:58 pm, "William Stein" <[EMAIL PROTECTED]> wrote:
> On Dec 4, 2007 6:54 AM, mabshoff
>
>
>
> <[EMAIL PROTECTED]> wrote:
> > On Dec 4, 3:32 pm, Burcin Erocal <[EMAIL PROTECTED]> wrote:
> > > Hello,
>
> > > On Tue, 4 Dec 2007 01:40:36 -0800 (PST)
>
> > > Alexander Dreyer <[EMAIL PROTECTED]> wrote:
> > > > The essential part of PolyBoRi (using the built-ininterface) can be
> > > > built in about 3 minutes on a Intel(R) Xeon(R) CPU5148 @ 2.33GHz
> > > > (using one cpu only). I'll try to find the corresponding scons
> > > > commands for the spkg and the Sage-wrapper and give to to Burcin.
>
> > > There is a new PolyBoRi package available at:
>
> > >http://www.risc.uni-linz.ac.at/people/berocal/sage/polybori-0.1-r3.spkg
>
> > > Changes are:
> > > - Alexander's changes to speed up the build process
> > > - Update to the latest CVS version
> > > - pass on MAKEOPTS to scons to allow parallel builds
>
> > > Building the package takes 4 mins 20 seconds on a single Intel(R)
> > > Pentium (R) D CPU 3.40GHz. Parallel make options (-jn) speed up the
> > > build as expected.
>
> > Excellent. Please also add new info to #469. What I would need is a
> > list of external dependencies for PolyBoRi so that I can update the
> > deps file.
>
> Nothing should go into the standard Sage until:
Well, the above didn't imply that it would ship as standard in 2.9,
but that we need to get all the infrastructure in place to get on the
way. Putting it into deps without building it per default is part of
the plan.
>
> (1) it has been an optional package for some nonzero amount of time,
>
> (2) enough people have had a chance to look over the code and
> try it in the context of Sage,
>
> (3) there is enough clear demand for it.
>
> Getting R into Sage is a high priority, but it is definitely more than
> a week away, since there isn't even an R optional package yet -- only
> an R experimental package -- which means that I know that the
> package fails to build on some platforms, etc., and hasn't been
> well tested. Regarding (2), (3), R is in very good shape.
>
> Atlas (automatically tuned linear algebra software) -- the fast basic
> numerical linear algebra library -- also satisfies (2), (3), but there
> are some build issues with it too.
>
> Regarding PolyBoRi, (1) it is now an optional package (I just added
> it to the optional package list). It would be good if everybody
> reading this with interest in polybori and some cycles to spare
> could type
> sage: install_package('polybori-0.1-r3')
> and report the results. Regarding (3) it's clear there is strong
> demand for PolyBoRi from several people. (2) needs to happen.
> Could Burcin post a link to his patch bundle that provides library
> support for PolyBoRi for Sage in this thread, so we can all try
> it out?
Yep, that should happen ASAP.
> Also, regarding (2) could somebody who isn't the author
> of PolyBori actually at least *skim through the code* and make
> some comments about the quality, maintainability, lack-of-bad
> stuff, etc.?
Malb? Could you comment on that?
> -- William
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~----------~----~----~----~------~----~------~--~---
