I was not doing a two symbolic matrix multiply. I was doing two, 400 x 400
numerical matrices created by genmatrix(). This is a pure theoretical
exercise. I don't know about applications. If there is one then it is
worth doing. If there isn?t, it is still worth doing because it might have
unknown uses. That happened to me when I was writing pw.mac. It had uses I
never thought of when I was writing it. I don?t know if matrix multiply has
uses for such large matrices. Maxima has the ability to do them just in
case I suppose.
Rich
-----Original Message-----
From: Raymond Toy
Sent: Saturday, June 09, 2012 5:04 PM
To: maxima at math.utexas.edu
Subject: Re: [Maxima] parallel cell executing and wxmaxima
On 6/9/12 9:27 AM, Richard Hennessy wrote:
> Hi:
>
> Along a similar line of reasoning. Could we make matrix multiplication
> in Maxima multi-threaded? This could produce a speed up of up to a
> factor of 8 on a i7 Intel processor. If the matrix is a matrix of
> functions then let the user worry about side effects of individual
> threads in multi-treading. That is not Maxima's problem, it is the
> users. I think as long as the user has the ability to turn off or turn
> on parallelization of code then there is no problem.
Umm, how big of a symbolic matrix are you trying to multiply? And if
it's that big where parallelizing it is worth doing, can you even
interpret the matrix in a meaningful way?
Numerical matrices are a different animal altogether, but maxima can't
parallelize that either in its current state since it's just using
(non-parallel) LAPACK, translated.
Ray
_______________________________________________
Maxima mailing list
Maxima at math.utexas.edu
http://www.math.utexas.edu/mailman/listinfo/maxima