[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1178318609.23795.214.camel@localhost.localdomain>
Date: Fri, 04 May 2007 15:43:29 -0700
From: Tim Chen <tim.c.chen@...ux.intel.com>
To: Christoph Lameter <clameter@....com>
Cc: "Chen, Tim C" <tim.c.chen@...el.com>,
"Siddha, Suresh B" <suresh.b.siddha@...el.com>,
"Zhang, Yanmin" <yanmin.zhang@...el.com>,
"Wang, Peter Xihong" <peter.xihong.wang@...el.com>,
Arjan van de Ven <arjan@...radead.org>,
linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: RE: Regression with SLUB on Netperf and Volanomark
On Fri, 2007-05-04 at 11:27 -0700, Christoph Lameter wrote:
>
> Not sure where to go here. Increasing the per cpu slab size may hold off
> the issue up to a certain cpu cache size. For that we would need to
> identify which slabs create the performance issue.
>
> One easy way to check that this is indeed the case: Enable fake NUMA. You
> will then have separate queues for each processor since they are on
> different "nodes". Create two fake nodes. Run one thread in each node and
> see if this fixes it.
I tried with fake NUMA (boot with numa=fake=2) and use
numactl --physcpubind=1 --membind=0 ./netserver
numactl --physcpubind=2 --membind=1 ./netperf -t TCP_STREAM -l 60 -H
127.0.0.1 -i 5,5 -I 99,5 -- -s 57344 -S 57344 -m 4096
to run the tests. The results are about the same as the non-NUMA case,
with slab about 5% better than slub.
So probably the difference is due to some other reasons than partial
slab. The kernel config file is attached.
Tim
View attachment "config-numa-slub" of type "text/plain" (25422 bytes)
Powered by blists - more mailing lists