xfs-masters
[Top] [All Lists]

[xfs-masters] xfs write speed regression 2.6.18.1 to 2.6.19.1

To: xfs-masters@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx
Subject: [xfs-masters] xfs write speed regression 2.6.18.1 to 2.6.19.1
From: "Mr. Berkley Shands" <bshands@xxxxxxxxx>
Date: Fri, 29 Dec 2006 07:37:31 -0600
Cc: Dave Lloyd <dlloyd@xxxxxxxxx>
Reply-to: xfs-masters@xxxxxxxxxxx
Sender: xfs-masters-bounce@xxxxxxxxxxx
User-agent: Thunderbird 1.5.0.9 (X11/20061222)
Write speeds have decreased 10% to 30% between 2.6.18.1 and 2.6.19.1.
Read speeds are unchanged at 1.15 GB/Sec.

Using a SuperMicro H8DC8 or Tyan 2895 or a Tyan 2915 (Socket F)
with 4 2.2GHz Opterons, 16GB RAM, dual LSI 8408E SAS controllers
into 4 X 4 Raid0s. XFS file system. The only difference is the Kernel Rev.
16 Seagate 7200.10 Sata drives, with 3.AAE firmware (very important!).
LSI 8408E firmware rev is 1.02.01-0158. Adapter readahead is disabled.
(Enabling readahead with this firmware costs 25% in write performance :-( )

Under 2.6.18.1, I/O peaks at 256.1 MB/Sec into each raid0 - 1GB/Sec.
The average is 230 MB/Sec over the first TB. With 2.6.19.1,
the peak is 220 MB/Sec, and the average is 170 MB/Sec.

EXT3 runs 2-3X slower than XFS for this benchmark, so it is hard
to see where the regression appeared. I'm not really too worried
about it, but that much of a decrease is worth reporting.
Since there were significant XFS changes between the revs,
it might not be worthwhile for me to chase the exact update
that causes this issue.

berkley


 
-- 

//E. F. Berkley Shands, MSc//

**Exegy Inc.**

3668 S. Geyer Road, Suite 300

St. Louis, MO  63127

Direct:  (314) 450-5348

Cell:  (314) 303-2546

Office:  (314) 450-5353

Fax:  (314) 450-5354

 

This e-mail and any documents accompanying it may contain legally 
privileged and/or confidential information belonging to Exegy, Inc.  
Such information may be protected from disclosure by law.  The 
information is intended for use by only the addressee.  If you are not 
the intended recipient, you are hereby notified that any disclosure or 
use of the information is strictly prohibited.  If you have received 
this e-mail in error, please immediately contact the sender by e-mail or 
phone regarding instructions for return or destruction and do not use or 
disclose the content to others.

 



[[HTML alternate version deleted]]


<Prev in Thread] Current Thread [Next in Thread>
  • [xfs-masters] xfs write speed regression 2.6.18.1 to 2.6.19.1, Mr. Berkley Shands <=