Date: Fri, 17 Sep 1999 11:17:48 -0700 (PDT) From: Matthew Dillon <dillon@apollo.backplane.com> To: Brad Knowles <blk@skynet.be> Cc: current@FreeBSD.ORG Subject: Re: 2xPIIIx450 results & NFS results (was More benchmarking stuff...) Message-ID: <199909171817.LAA54393@apollo.backplane.com> References: <XFMail.990917112639.lh@aus.org> <199909171658.JAA53751@apollo.backplane.com> <v0420553bb40826e849a4@[195.238.1.121]>
next in thread | previous in thread | raw e-mail | index | archive | help
: Might I then request that you help rewrite it so that it performs
:a much more comprehensive testing of OS/filesystem throughput?
:Myself, I'd really love to see something that lets you seriously
:stress your system along the lines of Greg Lehey's rawio, but instead
:at a higher level. IMO, bonnie sucks worse than postmark, although
:they're measuring different things.
:
: Although it should certainly be forking, whether forking or not I
:can tell you that creating huge directories is not necessarily a bad
:simulation of a heavily-used mail server. I've seen mail servers
:with over 100,000 files in /var/spool/mqueue, both at former
:employers (like AOL), and at former customer sites (such as some of
:the largest freemail providers in the world).
What we really need is something that generates a performance
curve based on several variables, including block size, locality of
reference (seek randomosity), amount of parallelism, locality of
parallelism (i.e. operating on same files vs different files), size of
dataset in bytes, and size of dataset in files.
The program should dynamically mess with all the variables until it
gets a statistically relevant curve.
I don't have the time to do it. Sniff!
-Matt
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199909171817.LAA54393>