Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-25 Thread Holger Hoffstätte
On 2020-06-25 12:42, Holger Hoffstätte wrote: Nevermind, it turned out to be something else after all. I'm not entirely sure *what* exactly yet, but for some reason my loadavg is high again for no good reason. Sorry for the false alarm. -h

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-25 Thread Vincent Guittot
On Thu, 25 Jun 2020 at 12:42, Holger Hoffstätte wrote: > > On 2020-06-25 11:56, Vincent Guittot wrote: > > On Thu, 25 Jun 2020 at 11:24, Holger Hoffstätte > > wrote: > >> > >> On 2020-06-24 17:44, Vincent Guittot wrote: > >>> Some performance regression on reaim benchmark have been raised with >

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-25 Thread Holger Hoffstätte
On 2020-06-25 11:56, Vincent Guittot wrote: On Thu, 25 Jun 2020 at 11:24, Holger Hoffstätte wrote: On 2020-06-24 17:44, Vincent Guittot wrote: Some performance regression on reaim benchmark have been raised with commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to classify

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-25 Thread Vincent Guittot
On Thu, 25 Jun 2020 at 11:24, Holger Hoffstätte wrote: > > On 2020-06-24 17:44, Vincent Guittot wrote: > > Some performance regression on reaim benchmark have been raised with > >commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to > > classify group") > > > > The problem comes

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-25 Thread Holger Hoffstätte
On 2020-06-24 17:44, Vincent Guittot wrote: Some performance regression on reaim benchmark have been raised with commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to classify group") The problem comes from the init value of runnable_avg which is initialized with max value. Thi

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-24 Thread Vincent Guittot
On Wed, 24 Jun 2020 at 18:32, Valentin Schneider wrote: > > > On 24/06/20 16:44, Vincent Guittot wrote: > > Some performance regression on reaim benchmark have been raised with > > commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to > > classify group") > > > > The problem comes

Re: [PATCH] sched/cfs: change initial value of runnable_avg

2020-06-24 Thread Valentin Schneider
On 24/06/20 16:44, Vincent Guittot wrote: > Some performance regression on reaim benchmark have been raised with > commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to > classify group") > > The problem comes from the init value of runnable_avg which is initialized > with max va

[PATCH] sched/cfs: change initial value of runnable_avg

2020-06-24 Thread Vincent Guittot
Some performance regression on reaim benchmark have been raised with commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to classify group") The problem comes from the init value of runnable_avg which is initialized with max value. This can be a problem if the newly forked task is