lkml.org 
[lkml]   [2011]   [Jun]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 7/7] memcg: proportional fair vicitm node selection
On Thu 23-06-11 23:10:11, Hiroyuki Kamezawa wrote:
> 2011/6/23 Michal Hocko <mhocko@suse.cz>:
> > On Thu 16-06-11 12:57:41, KAMEZAWA Hiroyuki wrote:
> >> From 4fbd49697456c227c86f1d5b46f2cd2169bf1c5b Mon Sep 17 00:00:00 2001
> >> From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> >> Date: Thu, 16 Jun 2011 11:25:23 +0900
> >> Subject: [PATCH 7/7] memcg: proportional fair node vicitm selection
> >>
> >> commit 889976 implements a round-robin scan of numa nodes for
> >> LRU scanning of memcg at hitting limit.
> >> But, round-robin is not very good.
> >>
> >> This patch implements a proportionally fair victim selection of nodes
> >> rather than round-robin. The logic is fair against each node's weight.
> >>
> >> Each node's weight is calculated periodically and we build an node's
> >> scheduling entity as
> >>
> >>      total_ticket = 0;
> >>      for_each_node(node)
> >>       node->ticket_start =  total_ticket;
> >>         node->ticket_end   =  total_ticket + this_node's_weight()
> >>         total_ticket = node->ticket_end;
> >>
> >> Then, each nodes has some amounts of tickets in proportion to its own weight.
> >>
> >> At selecting victim, a random number is selected and the node which contains
> >> the random number in [ticket_start, ticket_end) is selected as vicitm.
> >> This is a lottery scheduling algorithm.
> >>
> >> For quick search of victim, this patch uses bsearch().
> >>
> >> Test result:
> >>   on 8cpu box with 2 nodes.
> >>   limit memory to be 300MB and run httpd for 4096files/600MB working set.
> >>   do (normalized) random access by apache-bench and see scan_stat.
> >>   The test makes 40960 request. and see scan_stat.
> >>   (Because a httpd thread just use 10% cpu, the number of threads will
> >>    not be balanced between nodes. Then, file caches will not be balanced
> >>    between nodes.)
> >
> > Have you also tried to test with balanced nodes? I mean, is there any
> > measurable overhead?
> >
>
> Not enough yet. I checked OOM trouble this week :).
>
> I may need to make another fake_numa setup + cpuset
> to measurements.

What if you just use NUMA rotor for page cache?

> In usual path, new overhead is random32() and
> bsearch(). I'll do some.
>
> Thanks,
> -Kame

--
Michal Hocko
SUSE Labs
SUSE LINUX s.r.o.
Lihovarska 1060/12
190 00 Praha 9
Czech Republic
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2011-06-23 16:33    [W:0.082 / U:0.760 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site