From owner-freebsd-hackers  Tue Apr  1 05:47:00 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id FAA01988
          for hackers-outgoing; Tue, 1 Apr 1997 05:47:00 -0800 (PST)
Received: from shadows.aeon.net (bsdhack@shadows.aeon.net [194.100.41.1])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id FAA01976
          for <hackers@freebsd.org>; Tue, 1 Apr 1997 05:46:55 -0800 (PST)
Received: (from bsdhack@localhost)
          by shadows.aeon.net (8.8.5/8.8.3) id QAA20476
          for hackers@freebsd.org; Tue, 1 Apr 1997 16:47:25 +0300 (EET DST)
From: mika ruohotie <bsdhack@shadows.aeon.net>
Message-Id: <199704011347.QAA20476@shadows.aeon.net>
Subject: p166 vs. p166mmx
To: hackers@freebsd.org
Date: Tue, 1 Apr 1997 16:47:25 +0300 (EET DST)
X-Mailer: ELM [version 2.4ME+ PL31 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

well, i was curious about this, and wanted to find out the truth...

this is exactly identical set of src, though today's 2.1-STABLE since
my desktop at work runs stable, and i had to do it at work, couldnt
carry the cpu to home so i could've done it on 3.0-CURRENT.

machine was Gigabyte GA586HX512 board (with tag chip for full caching)
using 64 megs of true parity micron ram, in two 32 meg simms. it has
adaptec 2940 uw, only one quantum atlas-ii fast wide. i had /usr/src
and /tmp mounted async (noatime doesnt work in stable, i tried), /usr/obj
is goes to /tmp. oh, and ahc options scbpaging, tagenable and memio
were used. /tmp was empty in the beginning for both.

both created identical log from the compiling, i diffed it and only lines
that differed included echoed dates. it would've compiled without the -k
on both times.

i think i should mention machine was on net and i ran afterstep pre6 and
several xterms while it compiled.

first normal intel pentium 166mhz:
--
make -k world 7595.59s user 1843.94s system 87% cpu 3:00:48.28 total

make depend  25.66s user 9.29s system 79% cpu 44.132 total
text    data    bss     dec     hex
798720  53248   95220   947188  e73f4
make  237.70s user 27.34s system 84% cpu 5:12.39 total
960912 Apr  1 11:55
--

then, the intel pentium 166mhz with mmx
--
make -k world 5982.25s user 1636.02s system 87% cpu 2:24:21.88 total

make depend  21.54s user 8.26s system 78% cpu 37.911 total
text    data    bss     dec     hex
798720  53248   95220   947188  e73f4
make  188.30s user 25.57s system 89% cpu 3:58.24 total
960912 Apr  1 15:31
--

i was somewhat impressed, and surprised.


mickey