On Feb 1, 1:26 pm, William Stein <[email protected]> wrote:
<SNIP>
> > Maple's Linear Algebra is noted for not being particularly good at large
> > sizes. With symbolic entries, it's much worse. Thanks for these timings,
> > it just reinforces my idea that I should switch as soon as the calculus
> > improves.
>
> > Is there a particular reason why Magma is that much better for small
> > inputs?
>
> I would love to know. I've cc'd Allan Steel, since he would know.
> Maybe he'll tell us.
> Allan, for reference the complete thread for this discussion is:
>
> http://groups.google.com/group/sage-devel/browse_thread/thread/30d4a5...
>
> By the way, we are having major problems with Sage on Fedora 64-bit...
> and somebody mentioned you might have some tricks to fix this. The
> thread about that is here:
>
> http://groups.google.com/group/sage-devel/browse_thread/thread/2fb155...
I doubt that is the same problem the Magma people saw since that issue
seems related to randomization and/or prelinking. I have thought about
the problem we see some more and IMHO we should modify Singular to
ignore the alleged problem reported by calling sbrk() since high level
Singular does not care about memory allocations failing anyway since
it tends just to crash.
The other solution that works quickly is not to use mmap but the
system malloc for oMalloc allocations exposes another bug that needs
to be fixed anyway, but the way it looks that will be something scope
related in Cython, so I have no clue how to even approach this. My
guess would be that RobertWB would be the man who could fix this the
quickest, but he is busy with real life, so I am not sure how to
approach this.
> -- William
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
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://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
