Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-08 Thread Peter Zijlstra
On Tue, May 07, 2013 at 03:51:14PM +0800, Alex Shi wrote: > On 05/03/2013 03:55 PM, Alex Shi wrote: > > > >> That should probably look like: > >> > >>preempt_disable(); > >>raw_spin_unlock_irq(); > >>preempt_enable_no_resched(); > >>schedule(); > >> > >> Otherwise you might find a

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-07 Thread Alex Shi
On 05/03/2013 03:55 PM, Alex Shi wrote: > >> That should probably look like: >> >> preempt_disable(); >> raw_spin_unlock_irq(); >> preempt_enable_no_resched(); >> schedule(); >> >> Otherwise you might find a performance regression on PREEMPT=y kernels. > > Yes, right! > Thanks

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-03 Thread Alex Shi
On 05/03/2013 03:55 PM, Alex Shi wrote: > Just find a benchmark named as pthread_cond_broadcast. > http://kristiannielsen.livejournal.com/13577.html. will play with it. :) >> > I tried the pthread_cond_broadcast with/without my latest patchset, seems no clear performance change. without the pat

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-03 Thread Alex Shi
> That should probably look like: > > preempt_disable(); > raw_spin_unlock_irq(); > preempt_enable_no_resched(); > schedule(); > > Otherwise you might find a performance regression on PREEMPT=y kernels. Yes, right! Thanks a lot for reminder. The following patch will fix

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-02 Thread Peter Zijlstra
On Thu, May 02, 2013 at 08:38:17AM +0800, Alex Shi wrote: > On 3.8 kernel, the first problem commit is 5a505085f043 ("mm/rmap: > Convert the struct anon_vma::mutex to an rwsem"), It cause much > imbalance among cpus on aim7 benchmark. The reason is here, > https://lkml.org/lkml/2013/1/29/84. Hehe

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-01 Thread Alex Shi
On 05/01/2013 08:14 PM, Peter Zijlstra wrote: > On Sat, Apr 27, 2013 at 01:25:38PM +0800, Alex Shi wrote: >> This patchset bases on tip/sched/core. >> >> The patchset remove the burst wakeup detection which had worked fine on 3.8 > > Was this part of the original series from PJT or some patches af

Re: [PATCH v4 0/6] sched: use runnable load based balance

2013-05-01 Thread Peter Zijlstra
On Sat, Apr 27, 2013 at 01:25:38PM +0800, Alex Shi wrote: > This patchset bases on tip/sched/core. > > The patchset remove the burst wakeup detection which had worked fine on 3.8 Was this part of the original series from PJT or some patches afterwards? I missed a few months worth of patches so an

[PATCH v4 0/6] sched: use runnable load based balance

2013-04-26 Thread Alex Shi
This patchset bases on tip/sched/core. The patchset remove the burst wakeup detection which had worked fine on 3.8 kernel, since the aim7 is very imbalance. But rwsem write lock stealing enabled in 3.9 kernel. aim7 imbalance disappeared. So the burst wakeup care doesn't needed. It was tested on