> Is this a reasonable assumption? Should I expect to see a performance
> improvement, or would it be negligible?
>
> Any ideas or shared experience would be very much appreciated.
I have recently done comparison for applications of a difference stencil to
a 1000x1000 mesh. I have also written a short conference proceedings where
I suggest Common Lisp for scientific computing and report this result.
@inproceedings{NNeuss_2002b,
author = {N.~Neuss},
title = "On using {C}ommon {L}isp in scientific computing",
booktitle = "Proceedings of the CISC 2002",
journal = "LNCSE",
volume = "",
publisher = "Springer-Verlag",
year = "2003"
}
(You can get the preprint (which is identical to the proceedings version)
from <http://www.iwr.uni-heidelberg.de/sfb/PP/Preprint2002-40.ps.gz>. And
if someone is interested I'll put also the code somehwere on my homepage.)
Recently, I have improved this code (without understanding how:-), so that
the result reported there is slightly pessimistic. Now, I measure at least
a factor two improvement of CMUCL over dynamically coded C, namely:
C (dynamic) C (static) CMUCL
P2 (400 MHz) 4.3 sec 1.1 sec 1.5 sec
P4 (2.4 GHz) 0.59 sec 0.30 sec 0.28 sec
I would suspect that you might observe similar results.
Nicolas.
