Benchmarking data from a raid machine 2.6.34-rc3

Previous thread: [RFC][PATCH] Journal superblock update should send a barrier by Jan Kara on Thursday, April 22, 2010 - 5:25 pm. (3 messages)

Next thread: No one seems to be using AOP_WRITEPAGE_ACTIVATE? by Theodore Ts'o on Saturday, April 24, 2010 - 7:40 pm. (3 messages)
From: Keith Mannthey
Date: Friday, April 23, 2010 - 5:06 pm

At long last here is a good data set for filesystem comparison.  It is
from 2.6.34-rc3 (unpatched). It compares ext3 ext4 ext4-nobarrier,
btrfs, btrfs-nocow, xfs and jfs on a software raid setup with a decent
spindle count. 

http://btrfs.boxacle.net/repository/raid/2010-04-14_2004/2.6.34-rc3/2.6.34-rc3.html


As we know touring barriers off helps a lot with ext4 on this machine.
I think Steve Pratt send some notification to the btrfs folks. Feel free
to point others at this data.


A history from this setup can be found here:  (mostly btrfs related
data)

http://btrfs.boxacle.net/repository/raid/history/History.html


I hope to be do these large data set runs once in every major kernel
revision life cycle. 

Thanks,
  Keith Mannthey
  LTC FS-Dev 


--

Previous thread: [RFC][PATCH] Journal superblock update should send a barrier by Jan Kara on Thursday, April 22, 2010 - 5:25 pm. (3 messages)

Next thread: No one seems to be using AOP_WRITEPAGE_ACTIVATE? by Theodore Ts'o on Saturday, April 24, 2010 - 7:40 pm. (3 messages)