Re: librep's indirect threaded bytecode interpretter

[karlheg at bittersweet dot inetarena dot com (Karl M. Hegbloom)]

>>>>> "Michael" == Michael Livshin <mlivshin@bigfoot.com> writes:

    Michael> karlheg@bittersweet.inetarena.com (Karl M. Hegbloom) writes:
    >> I ran a simple benchmark today, between `rep'
    >> and `guile'.  I defined an iterative factorial function, (the classic
    >> one that passes an accumulator and (1- n) via a tail-call), and ran
    >> it 1000 times in a loop with (fact-it 1 400).  Guile took almost a
    >> minute to finish, and `rep' finished in about 5.2 seconds.

    Michael> hmmm...  actually, this looks more connected to the fact that Rep uses
    Michael> the GMP library and Guile doesn't.

 Yes, that's what John Harper, the primary author of `rep' says also.

>>>> "John" == John Harper <john@dcs.warwick.ac.uk> wrote:

    Karl> I ran a simple benchmark on `rep' and `guile' today...  I
    Karl> ran a `fact-it' (iterative tail-calling factorial, passing
    Karl> an accumulator and n) in a loop... 1000 (fact-it 1 400)'s.
    Karl> Guile took almost a minute, and Rep ran it in just over 5
    Karl> seconds.  The latest Rep is about 30 seconds faster than the
    Karl> version that I got from Debian

    John> cool :-)

    John> [ (fac 400) is quite a large number, so libgmp may be taking
    John> most of the strain here.. I typically see that (cvs) rep is
    John> at least twice as fast as guile on `normal' code ]