delorie.com/archives/browse.cgi | search |
X-pop3-spooler: | POP3MAIL 2.1.0 b 4 980420 -bs- |
Message-ID: | <19980629235144.61540@cerebro.laendle> |
Date: | Mon, 29 Jun 1998 23:51:44 +0200 |
From: | Marc Lehmann <pcg AT goof DOT com> |
To: | beastium <beastium-list AT Desk DOT nl> |
Subject: | bzip2-mx vs. bzip2-nomx (II) (and patch) |
Mail-Followup-To: | beastium <beastium-list AT desk DOT nl> |
Mime-Version: | 1.0 |
X-Operating-System: | Linux version 2.1.107 (root AT cerebro) (gcc version pgcc-2.91.43 19980628 (gcc2 ss-980502 experimental)) |
Status: | RO |
Lines: | 23 |
Now the picture is clearer: bzip2-nomx bzip2-mx slowdown pii 1.67 1.76 1.6% pentium-mmx 1.37 1.59 14% amdk6 2.27 2.67 15% cyrix6x86mx 1.75 2.22 21% intel rules :[]. I uploaded a patch against egcs-19980628 into ftp.goof.com:/pub/pcg/test/ (but 0621 is faster ;), for people to play (and hopefully improve), it adds these switches (among others): -mmx use mmx instructions, safe to mix fpu and mmx code -mmx-only use mmx instructions and assume there is no fpu around -----==- | ----==-- _ | ---==---(_)__ __ ____ __ Marc Lehmann +-- --==---/ / _ \/ // /\ \/ / pcg AT goof DOT com |e| -=====/_/_//_/\_,_/ /_/\_\ --+ The choice of a GNU generation | |
webmaster | delorie software privacy |
Copyright © 2019 by DJ Delorie | Updated Jul 2019 |