ECL? was ..Re: Runtime determination of share directories?

Michael Abshoff wrote:
>
> I do not see how this is relevant. There clearly seems to be demand for 
> hooking up external libraries into Maxima. I am pretty sure that if I 
> benchmark a some Lapack routines using ATLAS against the f2cl translated 
> Lapack in Maxima you would see a large performance difference. And I can 
>   
Numeric speed is not the reason to use maxima; there are many other
things that would be better suited.

 I have not heard any one complain about the lapack routines in maxima
being too slow.  In fact, I'm not sure any even uses them for anything. :-)

And one thing that we can do with the Lisp translation is massage it
relatively easily to do a bigfloat implementation.  (I haven't actually
done this, but I did hack something to support quad-doubles.  That
appeared to work for the couple of simple tests I tried.)  Perhaps
there's already an mpfr version of lapack.

> trivially build ATLAS to use 4 cores at the same time and BLAS level 3 
>   
Last time I tried to build ATLAS was not exactly trivial, and took the
better part of a day.  Perhaps it's better now.

Ray