Messages in this thread | | | Date | Thu, 04 Sep 2014 13:50:34 -0700 | From | Dave Hansen <> | Subject | Re: regression caused by cgroups optimization in 3.17-rc2 |
| |
On 09/04/2014 08:08 AM, Johannes Weiner wrote: > Dave Hansen reports a massive scalability regression in an uncontained > page fault benchmark with more than 30 concurrent threads, which he > bisected down to 05b843012335 ("mm: memcontrol: use root_mem_cgroup > res_counter") and pin-pointed on res_counter spinlock contention. > > That change relied on the per-cpu charge caches to mostly swallow the > res_counter costs, but it's apparent that the caches don't scale yet. > > Revert memcg back to bypassing res_counters on the root level in order > to restore performance for uncontained workloads.
A quick sniff test shows performance coming back to what it was around 3.16 with this patch. I'll run a more thorough set of tests and verify that it's working well.
| |