Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!noao!ncar!elroy.jpl.nasa.gov!sdd.hp.com!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!bloom-beacon!eru!hagbard!sunic!mcsun!hp4nl!swi.psy.uva.nl!rontelta
From: rontelta@swi.psy.uva.nl (Ronteltap Tamara)
Newsgroups: comp.lang.smalltalk
Subject: Smalltalk: RISC vs CISC
Keywords: smalltalk performace risc cisc
Message-ID: <6025@swi.swi.psy.uva.nl>
Date: 17 Jun 91 12:58:02 GMT
Organization: SWI, UvA, Amsterdam
Lines: 30

Hi,

I'm new to Smalltalk. I've just bought a ST book to aquaint myself with
OO programming (Savic). It looks like you there's a lot of productivity to
be gained.

Before I finish the book and start writing some programs myself. There's
this paper I have write to on computer architecture: RISC versus CISC.
(yeah, yeah, I know, boring and already decided, that's why I chose the
subject)

The paper is almost finished. I just need some performance figures for
a Smalltalk benchmark "with many small objects" on RISC and CISC machines.

I suppose SPARC and 80386 would be OK. I've read recent articles on the
Richards (sp?) OS benchmark, but didn't have the presense of mind to
save the articles. If I remember correctly, the numbers posted were from
different Smalltalk versions. The optimizations in these (I believe 2.5 and
4.0) made a *lot* of difference.

So, has anyone got some performance benchmark on a RISC and a CISC?
Preferably not a micro benchmark and preferably using the same ST versions.

If you can direct me to some (IEEE) paper on this, I would be very happy to.


Bye for now, hope to talk some real ST later.

Richard
(rontelta@fwi.uva.nl)