What do you mean about this data structure

Greetings!

Richard Fateman  writes:

> laurent couraud wrote:
> 
> 
> > Yes. For both. But maybe it is a bug in maxima. I try to solve a
> > system
> > Of equation and maxima take more than 450MB and stop with error.
> 
> Is it a system of linear equations (linear in the variables, but perhaps
> with non-numeric coefficients)  or is it a system of polynomials? Or something
> else?
> Do you have any reason to believe that the system could be solved with, say
> 10X as much storage?
> 
>   And this
> > After more than 15 hours of calculus.
> >
> >> Are you sure a more efficient representation of CREs will make a
> >> big difference?
> > I don't know. It is why I ask you.
> 
> The first thing to do is find out what is taking so long. There is an excellent
> profiling system in at least one implementation of Maxima (in Allegro
> Common Lisp)
> >

GCL also has two profiling modes, one available in all images (using
'profil'), and one more comprehensive option using gprof that is
available in images configured with this option.  Just bracket your
stuff between (si::gprof-start) and (si::gprof-quit).  The profiling
images are available on Debian by default, and are selectable for
example with

export GCL_PROF=t
GCL (GNU Common Lisp)  2.6.6 CLtL1  profiling  Jan 18 2005 00:06:53
Source License: LGPL(gcl,gmp), GPL(unexec,bfd)
Binary License:  GPL due to GPL'ed components: (READLINE BFD UNEXEC)
Modifications of this banner must retain notice of a compatible license
Dedicated to the memory of W. Schelter

Use (help) to get some basic information on how to use GCL.

>

Note the profiling in the header.  You can build maxima atop this and
profile to your heart's content.  

This is not the default in GCL as there is a performance penalty to
the -pg switch in gcc.  The penalty preserves proportionate times,
which is all that matters in this case.

Take care,


> 
> > I don't know because i don't know Lisp language. But maybe it is
> > possible
> > to do memory management for CRE independently of Lisp GC
> 
> Maybe you should learn lisp, because otherwise you will end up
> re-implementing 15% of it, but with bugs. This is the history
> of Mathematica, Maple, and other systems.
> 
> >
> >>Why do you have your own
> >>representation of the variables (x1 etc.) instead of using the
> >>existing representation?
> 
> The existing representation makes up new symbols (gensyms) for
> kernels.
> >
> 
> RJF
> 
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
> 
> 
> 

-- 
Camm Maguire			     			camm@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah