From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail143.messagelabs.com (mail143.messagelabs.com [216.82.254.35]) by kanga.kvack.org (Postfix) with SMTP id 8AFAC9000BD for ; Sun, 2 Oct 2011 08:47:27 -0400 (EDT) From: "Shi, Alex" Date: Sun, 2 Oct 2011 20:47:21 +0800 Subject: RE: [PATCH] slub Discard slab page only when node partials > minimum setting Message-ID: <6E3BC7F7C9A4BF4286DD4C043110F30B5FD97584A3@shsmsx502.ccr.corp.intel.com> References: <1315188460.31737.5.camel@debian> <1315357399.31737.49.camel@debian> <4E671E5C.7010405@cs.helsinki.fi> <6E3BC7F7C9A4BF4286DD4C043110F30B5D00DA333C@shsmsx502.ccr.corp.intel.com> <1315442639.31737.224.camel@debian> <1315557944.31737.782.camel@debian> <1315902583.31737.848.camel@debian> <1316050363.8425.483.camel@debian> <1316052031.8425.491.camel@debian> <1316765880.4188.34.camel@debian> <1317290032.4188.1223.camel@debian> In-Reply-To: Content-Language: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Sender: owner-linux-mm@kvack.org List-ID: To: Christoph Lameter Cc: Pekka Enberg , "Chen, Tim C" , "Huang, Ying" "Huang, Ying" , Andi Kleen , "linux-kernel@vger.kernel.org" , "linux-mm@kvack.org" > > I am tested aim9/netperf, both of them was said related to memory > > allocation, but didn't find performance change with/without PCP. Seems > > only hackbench sensitive on this. As to aim9, whichever with ourself > > configuration, or with Mel Gorman's aim9 configuration from his > > mmtest, both of them has no clear performance change for PCP slub. >=20 > AIM9 tests are usually single threaded so I would not expect any differen= ces. > Try AIM7? And concurrent netperfs? I used aim7+aim9 patch, and setup 2000 process run concurrently. But aim9=20 can't have big press on slab in fact.=20 As to concurrent netperf, I'd like try it after vacation, if you can wait. = :)=20 >=20 > The PCP patch helps only if there is node lock contention. Meaning > simultaneous allocations/frees from multiple processor from the same cach= e. >=20 > > Checking the kernel function call graphic via perf record/perf report, > > slab function only be used much in hackbench benchmark. >=20 > Then the question arises if its worthwhile merging if it only affects thi= s > benchmark. >=20 >>From my viewpoint, the patch is still helpful on server machines, while no = clear=20 regression finding on desktop machine. So it useful.=20 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org