On Fri, Aug 20, 2004 at 03:58:52PM -0700, Jason Howard wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> They are. On the SCSI setup we are using 2K SCSI transfers, so the data and
> log sectorsizes are set to 2K (or 4K).
> ...
> I tested Ext3 and JFS on the array we have here (the SCSI) and, while I'm not
> seeing as good write speeds as XFS, the read speeds are much better than the
> 67MB/s I was getting with XFS. JFS gives me about 150 MB/s on both the read
> and write. Ext3 give me slightly less.
It should be much closer than that - are you comparing XFS with
a 2K blocksize to the others with a 4K blocksize filesystem maybe?
The blocksize used makes a fairly large difference to the generic
IO path IIRC.
If you haven't done so already, you may want to also investigate if
the atime updates are having any affect (see "noatime" mount option)
on your numbers.
From my performance testing we are fairly close to the other fs's
on read performance, but slightly under. I haven't got to the
bottom of that yet, I don't know of any reason that we should be
slower there.
cheers.
--
Nathan
|