On 5/24/07, Robert Dodier <robert.dodier at gmail.com> wrote:
>
> On 5/23/07, Stavros Macrakis <macrakis at alum.mit.edu> wrote:
>
> > and it's really getting annoying that it insists on compiling the
> function...
> I'm guessing you have GCL + Maxima. GCL is much slower to compile
> functions because it must write to the disk.
Yes, and Andrej pointed out to me that the current version doesn't compile
the function under GCL.
Note, though, that in a nested integral, the current nintegrate re-compiles
the inner function each time the outer function evaluates it. Even if
compilation is faster than GCL, this is significant and unnecessary
overhead. There seem to be two possible ways to approach this:
-- have nintegrate be more clever about the nested case, but still use 1d
integrators
-- use some 2d integration code in quadpack, which I'm guessing exists
-s