From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755808Ab1KAAmi (ORCPT ); Mon, 31 Oct 2011 20:42:38 -0400 Received: from home.kolivas.org ([59.167.196.135]:34460 "EHLO mine.kolivas.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1755697Ab1KAAmh (ORCPT ); Mon, 31 Oct 2011 20:42:37 -0400 From: Con Kolivas To: Mike Galbraith Cc: "Artem S. Tashkinov" , linux-kernel@vger.kernel.org Subject: Re: HT (Hyper Threading) aware process scheduling doesn't work as it should Date: Tue, 01 Nov 2011 11:41:59 +1100 Message-ID: <2516934.OmMpfqGngS@quad> User-Agent: KMail/4.7.2 (Linux/2.6.38.4-ck1; KDE/4.7.2; x86_64; ; ) In-Reply-To: <1320061348.6389.18.camel@marge.simson.net> References: <269467866.49093.1320004632156.JavaMail.mail@webmail17> <1479015.f2tEjKr9vR@quad> <1320061348.6389.18.camel@marge.simson.net> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 31 Oct 2011 12:42:28 PM Mike Galbraith wrote: > On Mon, 2011-10-31 at 21:06 +1100, Con Kolivas wrote: > > On Sun, 30 Oct 2011 07:57:12 PM Artem S. Tashkinov wrote: > > > I've found out that even on Linux 3.0.8 the process scheduler > > > doesn't > > > correctly distributes the load amongst virtual CPUs. E.g. on a > > > 4-core > > > system (8 total virtual CPUs) the process scheduler often run some > > > instances of four different tasks on the same physical CPU. > > > > > > Any thoughts? comments? I think this is quite a serious problem. > > > > Intense cache locality logic, power saving concepts, cpu frequency > > governor behaviour and separate runqueues per CPU within the current > > CPU process scheduler in the current mainline linux kernel will > > ocasionally do this. Some workloads will be better, while others will > > be worse. Feel free to try my BFS cpu scheduler if you wish a CPU > > process scheduler that spreads work more evenly across CPUs. > > > > Alas the last version I synced up with will not apply cleanly past about > > 3.0.6 I believe: > > > > http://ck.kolivas.org/patches/bfs/3.0.0/3.0-sched-bfs-413.patch > > Yeah, it handles independent tasks well, but cache misses can be > excruciatingly painful for the others. > > Q6600 box, configs as identical as possible, tbench 8 > > 3.0.6-bfs413 728.6 MB/sec > 3.0.8 1146.7 MB/sec > > -Mike Fortunately BFS is about optimising user visible service latency for normal users running normal applications on normal desktops under normal workloads, and not about tbench throughput. Regards, Con -- -ck