Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1370839
| From | Rik van Riel <riel@redhat.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [RFC] sched: unused cpu in affine workload |
| Date | 2016-04-04 21:50 +0200 |
| Message-ID | <rkiFA-89A-5@gated-at.bofh.it> (permalink) |
| References | <rk83w-jm-13@gated-at.bofh.it> <rk8wx-tQ-1@gated-at.bofh.it> <rk8ZA-Wy-23@gated-at.bofh.it> <rk99g-118-21@gated-at.bofh.it> <rkcJP-3AF-3@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
[Multipart message — attachments visible in raw view] - view raw
On Mon, 2016-04-04 at 15:23 +0200, Peter Zijlstra wrote: > On Mon, Apr 04, 2016 at 11:38:44AM +0200, Ingo Molnar wrote: > > > > We'd upgrade that to O(nr_cpus^2), which is totally unrealistic > > with 16,000 CPUs > > even in a slowpath - but it would probably cause problems even with > > 120 CPUs. It > > will get quadratically worse as the number of CPUs in a system > > increases on its > > current exponential trajectory ... > The arbitrary affinity thing is I think a packing problem, which is > NP > hard IIRC. An optimal solution is NP hard. Heuristics that "move tasks with pressure" may be much more doable, and lead to perfectly satisfactory results, especially if most migrations happen within a socket (and the same shared L3 cache). -- All Rights Reversed.
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[RFC] sched: unused cpu in affine workload Jiri Olsa <jolsa@redhat.com> - 2016-04-04 10:30 +0200
Re: [RFC] sched: unused cpu in affine workload Peter Zijlstra <peterz@infradead.org> - 2016-04-04 10:50 +0200
Re: [RFC] sched: unused cpu in affine workload Ingo Molnar <mingo@kernel.org> - 2016-04-04 11:00 +0200
Re: [RFC] sched: unused cpu in affine workload Ingo Molnar <mingo@kernel.org> - 2016-04-04 11:30 +0200
Re: [RFC] sched: unused cpu in affine workload Ingo Molnar <mingo@kernel.org> - 2016-04-04 11:40 +0200
Re: [RFC] sched: unused cpu in affine workload Peter Zijlstra <peterz@infradead.org> - 2016-04-04 15:30 +0200
Re: [RFC] sched: unused cpu in affine workload Rik van Riel <riel@redhat.com> - 2016-04-04 21:50 +0200
Re: [RFC] sched: unused cpu in affine workload Peter Zijlstra <peterz@infradead.org> - 2016-04-04 23:40 +0200
Re: [RFC] sched: unused cpu in affine workload Jiri Olsa <jolsa@redhat.com> - 2016-04-05 11:00 +0200
csiph-web