delorie.com/archives/browse.cgi | search |
From: | sandmann AT clio DOT rice DOT edu (Charles Sandmann) |
Message-Id: | <10301100337.AA12120@clio.rice.edu> |
Subject: | Re: old packages on simtel |
To: | djgpp-workers AT delorie DOT com |
Date: | Thu, 9 Jan 2003 21:37:24 -0600 (CST) |
In-Reply-To: | <200301092315.h09NF3519198@envy.delorie.com> from "DJ Delorie" at Jan 09, 2003 06:15:03 PM |
X-Mailer: | ELM [version 2.5 PL2] |
Mime-Version: | 1.0 |
Reply-To: | djgpp-workers AT delorie DOT com |
Errors-To: | nobody AT delorie DOT com |
X-Mailing-List: | djgpp-workers AT delorie DOT com |
X-Unsubscribes-To: | listserv AT delorie DOT com |
> > ... gcc281?.zip ... > > Would you be receptive to adding those old packages if somebody can > > find them? > If there's a good reason to put the effort into them, sure. Gcc 2.7.x and gcc 2.8.1 seem to generate much smaller code on C source than 2.9.x or 3.x compilers. This has been attributed to everything from alignment to speed improvements, but for most benchmarks on utility code (with I/O, processing, not tight loops) I see little or no difference (in some cases the newer code is slower). The older compilers are more resource friendly in compile time and memory utilization. They also work fine with the newer binutils. The GCC developers seem more concerned with 1% speed improvements on some mythical benchmark on the latest chip than a 20% bloat in size (and slowdowns) for earlier architectures. I have a 2.8.1 gcc built with the v2.03 refresh someplace. I haven't had time to package it and make it available, but it's something I might consider someday if I could find the time. But right now it's a pretty low priority...
webmaster | delorie software privacy |
Copyright © 2019 by DJ Delorie | Updated Jul 2019 |