Re: Scheduler behaviour

2007-12-07 Thread Holger Wolf
Arjan van de Ven wrote: On Wed, 05 Dec 2007 21:15:30 +0100 Holger Wolf <[EMAIL PROTECTED]> wrote: We discovered performance degradation with dbench when using kernel 2.6.23 compared to kernel 2.6.22. In our case we booted a Linux in a IBM System z9 LPAR with 256MB of ram with 4 CPU's. This

Re: Scheduler behaviour

2007-12-06 Thread Jarek Poplawski
Arjan van de Ven wrote, On 12/05/2007 10:26 PM: > On Wed, 05 Dec 2007 21:15:30 +0100 > Holger Wolf <[EMAIL PROTECTED]> wrote: ... >> a 2.6.23 kernel. We saw a throughput degradation from 7.2 to 23.4 > > this is good news! > dbench rewards unfair behavior... so higher dbench usually means a > wo

Re: Scheduler behaviour

2007-12-05 Thread Arjan van de Ven
On Wed, 05 Dec 2007 21:15:30 +0100 Holger Wolf <[EMAIL PROTECTED]> wrote: > We discovered performance degradation with dbench when using kernel > 2.6.23 compared to kernel 2.6.22. > > In our case we booted a Linux in a IBM System z9 LPAR with 256MB of > ram with 4 CPU's. This system uses a stripe

Re: Scheduler behaviour

2007-12-05 Thread Ingo Molnar
* Peter Zijlstra <[EMAIL PROTECTED]> wrote: > > We discovered performance degradation with dbench when using kernel > > 2.6.23 compared to kernel 2.6.22. > > We've fixed a lot of regressions and made a lot of other changes to > the scheduler since .23. Could you please try the backport of the

Re: Scheduler behaviour

2007-12-05 Thread Peter Zijlstra
On Wed, 2007-12-05 at 21:15 +0100, Holger Wolf wrote: > We discovered performance degradation with dbench when using kernel 2.6.23 > compared to kernel 2.6.22. We've fixed a lot of regressions and made a lot of other changes to the scheduler since .23. Could you please try the backport of the l