[Maxima] is maxima suitable for multiprocessor computations?
mmarco at unizar.es
Tue Nov 13 10:26:06 CST 2007
El Martes, 13 de Noviembre de 2007 16:00, Stavros Macrakis escribió:
> Not likely. The algorithms in Maxima were not designed to be parallelized.
> The front ends (xmaxima etc.) and the plotting packages do run in separate
> processes, so in theory could benefit from more than one processor, but I
> doubt the benefit wo> On Nov 13, 2007 9:44 AM, Konovalov, Vadim
Vladimirovich (Vadim)** CTR ** <
> vkonovalov at alcatel-lucent.com> wrote:
> > I wonder whether Maxima could benefit of having more than 1 processor on
> > the system.uld be meaningful.
Maybe it would be a good idea to start considering the implementation of
parallel algorithms. Most of the processors shipped in new computers are
multicore, and it seems that in the future this tendency will increase more
Is there a structural limitation to the capability of maxima to run several
threads or the problem is just that the actual algorithms are not
parallelizable? If the issue is just the algorithms, maybe we could start by
just implementing a series of "paralezization tools" in order for the users
to program their own parallel algorithms. After that, the inclusion of
parallel algorithms for specific purposes could just be a matter of time and
re-using the work done by the users.
Miguel Marco Buzunariz
Departamento de Matematicas
Universidad de Zaragoza.
More information about the Maxima