[lkml]   [2011]   [Sep]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH v1] sched: fix nohz idle load balancer issues
* Srivatsa Vaddagiri <> [2011-09-28 10:04:27]:

> > Especially with higher number of logical CPUs. Likely cleaner way is to clear
> > first_pick, second_pick on idle instead of clearing on tickless.
> I think I tried that (cleared first/second_pick_cpu in
> nohz_kick_needed() upon idle) but didn't get the best results. Let me
> try that again and post idle time numbers.

So made some changes so that first/second_pick_cpu identifiers are
cleared when corresponding cpus become idle. They are cleared in
pick_next_task_idle() ..

With that change, I don't see additional benefit of having a third busy
cpu kick ilb_cpu when it has > 1 task.

Idle time -> Avg. Std.Dev Min Max
W/o third busy cpu kicking 4.57% 0.9 3% 8%
With third busy cpu kicking 4.78% 1.1% 3% 8%

So will go with your suggestion in the next version!

- vatsa

 \ /
  Last update: 2011-09-28 16:21    [W:0.093 / U:0.684 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site