Newsgroups: comp.lang.scheme,comp.lang.lisp,comp.lang.lisp.x
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!udel!gatech!swrinde!cs.utexas.edu!uwm.edu!lll-winken.llnl.gov!ames!kronos.arc.nasa.gov!ptolemy.arc.nasa.gov!kpc
From: k p c <kpc@ptolemy.arc.nasa.gov>
Subject: Benchmarking results for modern Lisps and Schemes
Message-ID: <1995Mar7.074834.18661@ptolemy-ethernet.arc.nasa.gov>
X-Disclaimer: No organization, company, or government is represented here.
X-Attribution: kpc
Lines: 57
Sender: usenet@ptolemy-ethernet.arc.nasa.gov (usenet@ptolemy.arc.nasa.gov)
Nntp-Posting-Host: phenotype.arc.nasa.gov
Reply-To: kpc@ptolemy.arc.nasa.gov
Cc: kpc@ptolemy.arc.nasa.gov
Organization: NASA Ames Research Ctr., AI Research and Aero. Facil. Branches.
	Disclaimer: neither a civil servant nor a representative.
Date: Tue, 7 Mar 1995 07:46:39 GMT
Xref: glinda.oz.cs.cmu.edu comp.lang.scheme:12261 comp.lang.lisp:16971 comp.lang.lisp.x:1440

I looked (*) for benchmarking results for recent versions of Lisp and
Scheme implementations on modern machines and did not find any.

Has anybody done informal or formal (**) speed or space tests for any
two of the following Lisps, preferably under Solaris (***)?

	emacs 19.2x
	xlisp-plus with no windows
	clisp
	allegro 4.2
	scheme48 or scsh
	scm or guile with slib
	stk with no windows
	elk with no windows
	vscm

Please reply by email if you can.  If I get useful results, I will
summarize in tabular form to comp.lang.lisp.  I am surprised that I
have not found a web page dedicated to Lisp performance results.

Please note that I am excluding several Lisps (****) and that I am
after effects that are greater than a factor of 2, such as might be
produced by image dumping or compilation.  I don't care about
differences in speed less than that, since other factors (in my case,
for the specific purpose of a scripting language (*****), CL
compatibility and the ability to parse with regexps) are more
important.

(*) Various web and ftp sites, comp.benchmarks, and in the Lisp- and
benchmarking-related FAQs.

(**) I would definitely be satisfied with times for the Gabriel Lisp
benchmarks (referenced in the MK FAQ) with as many variables as
possible controlled.

(***) Any other OS would be fine.

(****) I know about bigloo, cmucl, gcl, kcl, akcl, lucid, ilog talk,
jlisp, ylisp, siod, wcl, and harlequin.  Please know that they might
all be very good implementations or languages despite my not including
them above.  Allegro is included because I use it for other purposes
and therefore can use it as a point of reference if others comment.

Please know that I know about non-Lisp alternatives and reject them
for my purposes.  De gustibus.

(*****) The reader might have seen my post on the subject some time
ago.  Since then, scsh and guile have come out and I discovered
xlisp-plus.  None of those were reviewed in my post.  xlisp-plus
interests me the most of these three, despite the features of scsh and
guile, because it attempts slightly more to be compatible with CL.
But speed factors greater than 2x and lack of regexps might veto it.
I am not enough of a Schemer to benchmark Schemes, so I hope that
others will be able to comment.  I hope for a CLer response also.

---
kpc@ptolemy.arc.nasa.gov.  AI, multidisciplinary neuroethology, info filtering.
