Re: tiny problems

From: Johann Friedrich Heinrichmeyer <Fritz.Heinrichmeyer_at_FernUni-Hagen.de>
Date: Mon, 5 Aug 96 08:52:31 +0200

Lars Thomas Hansen writes:
>
> >Has anyone done any profiling? Some of the STk examples
> >are slow in places. I'm wondering if it's Tk, Scheme,
> >or, probably, both.
>
> Performance-wise, the Scheme part of STk is a dog; my rule of thumb is a
> factor of 100 slower than native code on programs that do a lot of
> procedure calls and integer arithmetic (standard stupid benchmark to
> substantiate this is the doubly recursive Fibonnaci function). There


A factor 50 .. 100 slower than native code is probably the fastest You can
get from an interpreter with Your mentioned examples in any language
..

My impression was that SCM from A. Jaffer is the fastest scheme with
this examples, about twice as fast than Snow and also faster than
mzscheme. Strange, but snow was faster than siod. On the other hand
test.scm or r4rstest.scm from the SCM distribution and a lot of small
"throw away scripts" (i use Scheme or even emacs-lisp instead of Excel
for casual calculations ...) showed no significant
difference. Sometimes Snow even was faster in terms of "time .."
execution.

Maybe the relative low speed with recursive functions has to do with
the debugging feature of stk ..?
Received on Mon Aug 05 1996 - 08:55:58 CEST