cacheing compiled integrand in quadpack functions



On 1/11/06, Raymond Toy <raymond.toy at ericsson.com> wrote:

> I believe the external process only happens with gcl.  I think every
> other Lisp we support has the compiler built-in so compiling is very
> fast compared to gcl, even for cmucl which has a relatively slow
> compiler.

It's too bad gcl needs to hit the disk, since it is otherwise substantially
faster than sbcl and cmucl and much faster than clisp,
so far as I can tell.

Compiling the quadpack integrand vs not compiling it seems to yield
about the same overall elapsed time, in some limited tests I did,
so I'm going to just leave it be.

Robert