Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.wildstar.net!news.ececs.uc.edu!newsfeeds.sol.net!news.maxwell.syr.edu!news1.best.com!nntp1.ba.best.com!not-for-mail
From: dillon@flea.best.net (Matt Dillon)
Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.bsdi.misc,comp.sys.sgi.misc
Subject: Re: no such thing as a "general user community"
Date: 28 Mar 1997 11:30:22 -0800
Organization: BEST Internet Communications, Inc.
Lines: 73
Message-ID: <5hh68e$a8a@flea.best.net>
References: <331BB7DD.28EC@net5.net> <5hd29s$e7t@fido.asd.sgi.com> <5heptl$8ir@flea.best.net> <5hflpt$a3t@fido.asd.sgi.com>
NNTP-Posting-Host: flea.best.net
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:37951 comp.unix.bsd.bsdi.misc:6488 comp.sys.sgi.misc:29488
:In article <5hflpt$a3t@fido.asd.sgi.com>,
:Larry McVoy <lm@slovax.engr.sgi.com> wrote:
:>Matt Dillon (dillon@flea.best.net) wrote:
:>: :In article <5hd29s$e7t@fido.asd.sgi.com>,
:>: :Larry McVoy <lm@slovax.engr.sgi.com> wrote:
:>: :...
:...
:>: :>Larry McVoy lm@sgi.com http://reality.sgi.com/lm (415) 933-1804
:>
:>: IRIX scale to 100s of cpus ? Not in its current incarnation.
:>
:>You are right, in the 6.4 release it falls over at about 64 cpus. 6.5
:>fixes that.
:>
:>: There
:>: is nothing magical about XFS and there is nothing magical about striping
:>
:>It's magical only in that we do it well. It's fine to say that it is no big
:>deal. It actually is a big deal. It's hard to do it well and if it wasn't
:>hard how come FreeBSD doesn't do it?
:>
I think you guys need a reality check. Try actually running some of these
configurations on heavily loaded production machines rather then ad hoc
test beds or relatively simple configurations that are running only a few
distinct programs. This is about the umpteenth time I've heard this from
an SGI employee, but every time I dig it turns out that their miracle
machine isn't actually running a real load. Instead, it's running a
few 'test' suites. Bzzzttt... I can do that to, and get phenominal
results but you will never see me actually use those numbers for anything
beyond the simplest of comparisons.
:>
:>I'm working on a system that will be sustaining over 1GB/sec of NFS I/O.
:>That's 1GB/sec through the file system and the network. Oh, and just so
:>you don't think that's where it stops. There are over 20 of these suckers
:>running 24/7. It's around 28GB/sec of NFS/BDS/XFS I/O, sustained. Oh,
:>yeah, did I mention that they are all randoms? No sequential?
:>
:>Check back with me when you can play with the big boys.
That's fine, you guys can go off and play with the big boys... what few
of them there are, it makes your business really nice and, uh, unstable.
But hey, if you are willing to drop 95% of your potential market to
play with the 'big boys', please be my guest.
:>: IS something to be said for keeping utility programs up to date, fixing
:>: serious bugs, and keeping the friggin kernel up to date.
:>
:>We're working on it. I think we'll have an answer on the utilities that will
:>make you happy in about 9 months. It includes you getting source.
Right direction, wrong time frame.
:>We have a lot of people here that actually want to make things better.
:>I'll make you a challenge: you come up with a test case that reproduces
:>some bad ju ju in IRIX that works fine in FreeBSD. For each of those,
:>we'll fix it or tell you why it can't be fixed. I'll personally buy
:>you dinner or whatever for each of the ones that we can't fix just so
:>you aren't wasting your time. Don't like those rules? Come up with
:>some you do. I'm willing to play.
:>--
:>---
:>Larry McVoy lm@sgi.com http://reality.sgi.com/lm (415) 933-1804
This is about the third or fourth time someone at SGI has offered this
to us. We've taken them up on it in the past, and it hasn't panned out.
I, personally, do not have the time or inclination to keep on doing it
like a broken record so I will respectfully decline the offer.
-Matt