[634] in Release_7.7_team
Re: Athena 8.0 is slow on Solaris
daemon@ATHENA.MIT.EDU (John Hawkinson)
Mon Jul 8 21:08:11 1996
Date: Mon, 8 Jul 1996 21:08:04 -0400
To: Bill Cattey <wdc@MIT.EDU>, Greg Hudson <ghudson@MIT.EDU>
Cc: release-team@MIT.EDU, testers@MIT.EDU, sipb-staff@MIT.EDU
In-Reply-To: "[3000] in testers"
From: John Hawkinson <jhawk@MIT.EDU>
FYI, I rebooted bart-savagewood and re-ran my tests.
Typescript is in ~jhawk/80bench/8.0-beta-savagewood3.
Numbers:
configure: 25.71u 38.99s 1:46.06 61.0%
gmake: 204.44u 70.16s 5:00.59 91.3%
This is what I would expect, and in facts represents a speedup from
7.7 on bill-the-cat:
configure: 24.09u 45.79s 1:54.10 61.2%
gmake: 201.55u 108.42s 5:43.23 90.3%
This is without taking the update on bart.
I have now updated savagewood to 8.0E; Note to sipb-staff, it is now
AUTOUPDATE=true with appropriate mechanisms in /var/server/.private to
correct local customizations such that this should be a safe condition
for the machine to be in.
It is not clear to me whether the fact that rebooting fixed this is a
good thing or a bad thing, but it does suggest that this is not a
simple problem, and I will try and make sure that bart is doing a lot
of resource intensive tasks for the next few hours, in hopes that this
will bring out whatever sort of bugs.
Other SIPB folks should try and pound on bart if possible.
Thanks.
--jhawk
ps: I'll rerun my tests under 8.0E and see if there's a change.
Script output to go in ~jhawk/80bench/8.0-bart-savagewood3.