Disk performance in phoronix

Gilboa Davara gilboad at gmail.com
Mon Jul 27 09:37:39 UTC 2009


On Sun, 2009-07-26 at 21:24 -0500, T. Howell-Cintron wrote:
> T. Howell-Cintron wrote:
> > I'm stunned by the results and though I'm no guru I suspect SELinux
> > might have something to do with it.  What can be done to achieve better
> > - hopefully comparable - performance with Fedora?
> It's also troubling that the disk I/O, computational performance, and
> more seem to be slower as new releases are made available.  They
> benchmarked F7 through F10 and the different was sometimes dramatic.  We
> call that progress?
> 
> -- Tom 
> 
 
As far as I understand, the main problem with Phoronix's test suite is
that it doesn't use native packages.
Sure, Phoronix' copy of bzip2/apache/etc might be slower on Fedora 11
compared to Fedora 8, but it more-or-less says -zero- about the actual
performance difference between the -native- versions of bzip2/apache/etc
on Fedora 8 and Fedora 11. 

I'm not saying that Phoronix is wrong - I am saying that his testing
methodology is invalid.

- Gilboa




More information about the fedora-list mailing list