Path: utzoo!attcan!uunet!cs.utexas.edu!csd4.csd.uwm.edu!uxc.cso.uiuc.edu!uxc.cso.uiuc.edu!m.cs.uiuc.edu!p.cs.uiuc.edu!johnson
From: johnson@p.cs.uiuc.edu
Newsgroups: comp.lang.smalltalk
Subject: Re: Smalltalk for Scientific Applicatio
Message-ID: <80500068@p.cs.uiuc.edu>
Date: 16 Aug 89 16:10:00 GMT
References: <14@<Aug>
Lines: 16
Nf-ID: #R:<Aug:14:p.cs.uiuc.edu:80500068:000:838
Nf-From: p.cs.uiuc.edu!johnson    Aug 16 11:10:00 1989


While user primitives are the best way at present to make a Smalltalk
program fast, they are not the ultimate long-term solution.  What is
really needed is an optimizing compiler for Smalltalk.  My optimizing
compiler is not yet usable by people outside our research group, but
it makes Smalltalk programs an order of magnitude faster, will compile
all the usual Smalltalk programs, works with a modified version of
the Smalltalk debugger, and in general should produce an environment
that is just like Smalltalk-80 except with programs as fast as those
written in C.

Programs that have a lot of polymorphism cannot be optimized as much,
but a FORTRAN program that has been recoded in Smalltalk and compiled
with our compiler should be the same speed as the original program.  

Ralph Johnson - University of Illinois at Urbana-Champaign