Subject: Initial- and boundary-value problems in Maxima.
From: Richard Fateman
Date: Wed, 14 Sep 2011 16:30:15 -0700
On 9/14/2011 4:18 PM, Raymond Toy wrote:...
>
> I don't know why it's 100 times slower, though. Perhaps some
> additional optimizations by f2cl will reduce the cost.
I was puzzling over this too. I would expect that with sufficient
cleverness (which might entail writing some pieces of code directly in
lisp) one could get within 2X or so of fortran.... or -- for those
lisps that support it -- directly calling via a foreign function
interface, the fortran code would be just as fast.
If speed is important, the next step is, in my opinion, running some
benchmarks with profiling so we know where the time is spent. Is it
actually in the array-referencing (are we calling subroutines to compute
the array indexes? the code that is output from f2cl is not easy to
figure out). Or is it in allocating floating point number "boxes" ? Or
something else?
>
> The nice thing about the conversion is that with a little bit of work,
> we can actually have a bigfloat version (that will run even slower!).
>
> Ray
>
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima