[3000] in testers
Re: Athena 8.0 is slow on Solaris
daemon@ATHENA.MIT.EDU (John Hawkinson)
Mon Jul 8 17:14:00 1996
Date: Mon, 8 Jul 1996 17:13:35 -0400 (EDT)
To: Bill Cattey <wdc@MIT.EDU>
Cc: Greg Hudson <ghudson@MIT.EDU>
release-team@MIT.EDU, testers@MIT.EDU
In-Reply-To: "[628] in Release_7.7_team"
From: John Hawkinson <jhawk@MIT.EDU>
Perhaps this is all moot now, but just to tidy up your questions,
I provide answers.
> Could you please give me more details of your environment so that the
> tests you profess to have made can be repeated by me.
Sure; note that I said:
} Typescripts of both sessions are in ~jhawk/80bench.
they were intended to be there for clarity, and use by others in
reproducing; the "sequence of commands" was just to provide an
overview.
> 1. Stat should be stat.
Stat was a reference to stat(); I merely used ls -l.
> 2. I had to remove configure.log since I didn't have permission to bash it.
??
> 3. fs flushv returns an error of invalid argument when given /srvd and /os
Yeah, I fed it the output of ls -l.
> 4. ./gmake returns:
>
> ./gmake: Command not found.
An error on my part, was just "gmake".
In my original mail I said:
> bart-savagewood, a ??mb SPARC Classic running 8.0B
That should say 24.
On to your 2nd message:
> If I assume that by "time ./gmake" you mean "time gmake" and if I
> assume you mean I should build tcpdump not libpcap (since the latter
> builds in under two minutes.)
Yes, that's correct; the libpcap symlink is there so tcpdump
can find it to link with.
> hanta-yo% time gmake
> 214.40u 104.18s 6:47.32 78.2%
Great!
> I WILL THANK YOU TO BE very clear IN FUTURE WHEN YOU CHOOSE TO DROP
> IN AT THE LAST MINUTE WITH UNSUBSTANTIATED AND UNREPRODUCABLE
> PROBLEMS WITH MY RELEASE.
My message provided reference to where you could find sufficient
detail to accurately reproduce (the typescripts).
> I AM GRATEFUL FOR YOUR HELP, BUT PLEASE DO SO IN A TIMELY AND WELL
> ORGANIZED MANNER FROM NOW ON.
Greg solicited benchmarks and I responded to his solicitation in a
timely fashion; I regret that I haven't had time to pound on the beta
release very much at all. I was focusing on producing clear results,
though it's true that my mail reporting them could have been better
organized.
> There is no performance problem in 8.0 relative to 7.7.
I certainly hope that's the case. It's clear that there's some sort of
problem with bart-savagewood that does not exist on hanta-yo;
hopefully it means bart-savagewood has some problem that is unique to
it, but I would be reluctant to assume that without more evidence...
I will attempt to gather more information tonight.
--jhawk
