Newsgroups: comp.lang.prolog
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!news.kei.com!hookup!swrinde!cs.utexas.edu!uunet!allegra!ulysses!alice!pereira
From: pereira@alta.research.att.com (Fernando Pereira)
Subject: Re: Prolog benchmarking... how ?
In-Reply-To: conway@mundil.cs.mu.OZ.AU's message of Thu, 15 Sep 1994 22:46:56 GMT
Message-ID: <PEREIRA.94Sep15221606@alta.research.att.com>
Sender: netnews@ulysses.homer.att.com (Shankar Ishwar)
Reply-To: pereira@research.att.com
Organization: AT&T Bell Laboratories
References: <34qvro$6f3@tribune.usask.ca> <PEREIRA.94Sep10191603@alta.research.att.com>
	<ryoung.55.0009E235@utdallas.edu> <9425908.800@mulga.cs.mu.OZ.AU>
Date: Fri, 16 Sep 1994 02:16:06 GMT
Lines: 28

In article <9425908.800@mulga.cs.mu.OZ.AU> conway@mundil.cs.mu.OZ.AU (Thomas Charles CONWAY) writes:
> ryoung@utdallas.edu (Young U. Ryu) writes:
> >[Fernando Pereira talked about "bench_mark" contributed by many
> > well-known People.]
> >It would be great to put this in FAQ ...
> >Also, it would be great to put the benchmark results of
> >Prolog interpreters and compilers (either freeware or commercial).
> >If possible, I can volunteer...
> Unfortunately, it doesn't necessarily mean much to look at the
> benchmarks of language implementations carried out on any machine
> but your own. Here at Melbourne Uni, we have been benchmarking
> our new compiler for a pure logic language, and we get some very
> weird anomolies between machines. The causes of these strange
> behaviours (such as things getting slower when we use machine
> registers for storage instead of global variables) are probably
> due to caching effects, and possibly in some cases pipelining and
> super-scalarity effects.
Sigh, how life has become complicated since the days of the DEC-10, or
even the VAX or the MC68010... You are absolutely right, execution unit
concurrency and memory hierarchy are making simple benchmarks like
those I mentioned far too unreliable a predictor of performance in
actual applications.
--
Fernando Pereira
2D-447, AT&T Bell Laboratories
600 Mountain Ave, PO Box 636
Murray Hill, NJ 07974-0636
pereira@research.att.com
