Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Jul 2005 19:00:48 +1200
From:      Mark Kirkwood <markir@paradise.net.nz>
To:        freebsd-stable@freebsd.org
Subject:   FreeBSD IO Performance (was Re: Quality of FreeBSD)
Message-ID:  <42E099A0.3080101@paradise.net.nz>
In-Reply-To: <1dbad31505072105401c06bee6@mail.gmail.com>
References:  <1dbad31505072105401c06bee6@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
I happened to have received a 'new' machine, and wanted to see what its
IO system was capable of. So took the opportunity to run 4.10 and 5.4
against each other a few times. (fresh re-installs each time).

Its documented at:

http://homepages.paradise.net.nz/markir/freebsd/

I wanted to play with Docbook as well :-), so excuse the "book" format
(might be a few typos too).

But to cut to the chase, the results were overall very similar - 4.10
probably a little (4-8%) faster (allowing for run variation). So really
  5.4 is reasonably fast. The actual figures weren't too bad either -
70-80Mb/s read and writes on a 2 disk ATA array.

The most interesting thing discovered, was 5.4's "out of the box"
sequential *read* performance was considerably less then 4.10, but could
be brought up to almost the same by setting.

vfs.read_max=16

Hope this provides some interest, again - gotta qualify, this is all one
man's experiment on his hardware...

Cheers

Mark

P.s : of course, it would be nice if 5.x (or perhaps more importantly
6.x) was *faster* than 4.10....

Michael Schuh wrote:
> Hi,
> 
> Now my question to you : is the performance of ata-related disk-access
> under UFS-Filesystem not important for other application, so that the
> performance can be a half of them that RELENG_4 does?
> 
> In fact under RELENG_4 i can write a GIG FIle double as fast as under
> RELENG_5 ! and i would not hear any thing about serial performance or
> that this is not really like the real world, if i syimulate that with:
> 
> /usr/bin/time dd if=/dev/zero of=/zerofile bs=1024 count=1024k;
> this is reality poor!
> 






Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42E099A0.3080101>