Oliver Jehle wrote:
>
> hi ...
>
> i've read an article from christoph hallwig in a german magazine (Linux
> Magazin 1/2002) and see there, that jfs performance in the postmark
> benchmark is very bad. i have done some tests for my own with
> postmark ( http://people.redhat.com/arjanv/postmark-1_5RH.c ) because i
> can't trust what i've read.
>
> i've used a pentium III/700 with 512 MB RAM with kernel 2.4.17 / jfs
> 1.0.12 / xfs 1.0.2 and a seperate partition on a ide disk. i've always
> done a mkfs before running postmark. i've not taken a lot of time to
> look for the values representing the typical file system use or file
> sizes, i've done only a test to check if the values can be true that
> i've read.
>
> and my results for jfs arn't also are not very good in contrast to
> reiserfs/xfs.
>
> i would ask in the round, if there are some changes in the pipeline
> improving the performance to a level of XFS or Reiserfs ???
Thanks for the time you have taken to run postmark on the various
journaling file systems. We will take a look and see while JFS isn't
performing as good as it should be in this benchmark.
Thanks,
Steve
_______________________________________________
Jfs-discussion mailing list
[EMAIL PROTECTED]
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion
