Newsgroups: comp.lang.lisp
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!EU.net!uknet!festival!edcogsci!jeff
From: jeff@aiai.ed.ac.uk (Jeff Dalton)
Subject: Re: Fannkuch revisited or Benchmarking is hard
Message-ID: <Cx5s40.3nx@cogsci.ed.ac.uk>
Sender: usenet@cogsci.ed.ac.uk (C News Software)
Nntp-Posting-Host: bute.aiai.ed.ac.uk
Organization: AIAI, University of Edinburgh, Scotland
References: <Cwqt17.KA8@aplcenmp.apl.jhu.edu> <CwuKF8.F61@cogsci.ed.ac.uk> <36ck20$hqd@ursula.ee.pdx.edu>
Date: Tue, 4 Oct 1994 17:19:12 GMT
Lines: 31

In article <36ck20$hqd@ursula.ee.pdx.edu> marcus@ee.pdx.edu (Marcus Daniels) writes:
>jeff@aiai.ed.ac.uk (Jeff Dalton) writes:
>
>>A common view seemed to be that it would
>>all be too implementation-specific.  I disagree, but it's worth doing
>>even if I'm wrong. 
>
>How about starting with some examples of what some of these
>implementation-specific things are, and why complete dpANS declarations 
>aren't sufficient.

I don't have any good implementation-specific examples handy (remember
it's not *my* view that it's all too implementation-specific), but I
can recall doing the following for a numeric / array benchmark:

#+:cmu (deftype index () `(unsigned-byte 32))

#-:cmu (deftype index () 'fixnum)

(It was a Lisp vs Fortran matrix multiply benchmark posted by
Fateman a while back.)

Anyway, to get the best performance from CMU CL, I had to use
(unsigned-byte 32) as the array index type.  I'm not sure how
many other Common Lisps would handle that as well as (or better
than) fixnum.

I'm not sure what you mean by "complete dpANS declarations".
Which types would you pick?  Where would you use "the"?

-- jeff
