?iga Lenar?i? wrote:
>
> So flonum is always 'deftyped' to some internal floating point type.
> In a C program the situation is very clear - deftypes are resolved
> before compilation by the macro preprocessor, so optimization is there.
I don't think that's true. The C preprocessor runs first, and it
usually doesn't know anything about typedef. But that's off topic.
> Perhaps we could also support http://common-lisp.net/project/oct/
> which should work on all lisps, though it's quite a performance hit
> and I don't know if it's doable or even worth it...
It's certainly doable, but it depends on having a good compiler with a
fairly complete implementation of ANSI CL. Clisp works. (But then
clisp has long-float, so what's the point?) I'm pretty sure gcl won't
work because it's missing a lot of stuff. Maybe the missing stuff can
be ignored.
But I certainly wouldn't want maxima to use quad-doubles everywhere
since they are really quite slow. Double-doubles are also rather slow,
so I never compile maxima with double-double support.
Ray