From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Zhang, Yanmin" Subject: Re: Mainline kernel OLTP performance update Date: Thu, 12 Feb 2009 13:22:33 +0800 Message-ID: <1234416153.2604.387.camel@ymzhang> References: <1232616517.11429.129.camel@ymzhang> <1232617672.14549.25.camel@penberg-laptop> <1232679773.11429.155.camel@ymzhang> <4979692B.3050703@cs.helsinki.fi> <1232697998.6094.17.camel@penberg-laptop> <1232699401.11429.163.camel@ymzhang> <1232703989.6094.29.camel@penberg-laptop> <1232765728.11429.193.camel@ymzhang> <84144f020901232336v71687223y2fb21ee081c7517f@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Christoph Lameter , Andi Kleen , Matthew Wilcox , Nick Piggin , Andrew Morton , netdev@vger.kernel.org, Stephen Rothwell , matthew.r.wilcox@intel.com, chinang.ma@intel.com, linux-kernel@vger.kernel.org, sharad.c.tripathi@intel.com, arjan@linux.intel.com, suresh.b.siddha@intel.com, harita.chilukuri@intel.com, douglas.w.styner@intel.com, peter.xihong.wang@intel.com, hubert.nueckel@intel.com, chris.mason@oracle.com, srostedt@redhat.com, linux-scsi@vger.kernel.org, andrew.vasquez@qlogic.com, anirban.chakraborty@qlogic.com, Ingo Molnar To: Pekka Enberg Return-path: Received: from mga10.intel.com ([192.55.52.92]:63638 "EHLO fmsmga102.fm.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750785AbZBLFWu (ORCPT ); Thu, 12 Feb 2009 00:22:50 -0500 In-Reply-To: <84144f020901232336v71687223y2fb21ee081c7517f@mail.gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: On Sat, 2009-01-24 at 09:36 +0200, Pekka Enberg wrote: > On Fri, 2009-01-23 at 10:22 -0500, Christoph Lameter wrote: > >> No there is another way. Increase the allocator order to 3 for the > >> kmalloc-8192 slab then multiple 8k blocks can be allocated from on= e of the > >> larger chunks of data gotten from the page allocator. That will al= low slub > >> to do fast allocs. >=20 > On Sat, Jan 24, 2009 at 4:55 AM, Zhang, Yanmin > wrote: > > After I change kmalloc-8192/order to 3, the result(pinned netperf U= DP-U-4k) > > difference between SLUB and SLQB becomes 1% which can be considered= as fluctuation. >=20 > Great. We should fix calculate_order() to be order 3 for kmalloc-8192= =2E > Are you interested in doing that? Pekka, Sorry for the late update. The default order of =EF=BB=BFkmalloc-8192 on 2*4 stoakley is really an= issue of calculate_order. slab_size order name ------------------------------------------------- 4096 3 sgpool-128 8192 2 kmalloc-8192 16384 3 kmalloc-16384 =EF=BB=BFkmalloc-8192's default order is smaller than =EF=BB=BFsgpool-1= 28's. On 4*4 tigerton machine, a similiar issue appears on another kmem_cache= =2E =46unction =EF=BB=BFcalculate_order uses 'min_objects /=3D 2;' to shrin= k. Plus size calculation/checking in slab_order, sometimes above issue appear. Below patch against 2.6.29-rc2 fixes it. I checked the default orders of all kmem_cache and they don't become sm= aller than before. So the patch wouldn't hurt performance. =EF=BB=BFSigned-off-by Zhang Yanmin --- diff -Nraup linux-2.6.29-rc2/mm/slub.c linux-2.6.29-rc2_slubcalc_order/= mm/slub.c --- linux-2.6.29-rc2/mm/slub.c 2009-02-11 00:49:48.000000000 -0500 +++ linux-2.6.29-rc2_slubcalc_order/mm/slub.c 2009-02-12 00:08:24.00000= 0000 -0500 @@ -1856,6 +1856,7 @@ static inline int calculate_order(int si min_objects =3D slub_min_objects; if (!min_objects) min_objects =3D 4 * (fls(nr_cpu_ids) + 1); + min_objects =3D min(min_objects, (PAGE_SIZE << slub_max_order)/size); while (min_objects > 1) { fraction =3D 16; while (fraction >=3D 4) { @@ -1865,7 +1866,7 @@ static inline int calculate_order(int si return order; fraction /=3D 2; } - min_objects /=3D 2; + min_objects --; } =20 /*