Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1422635AbWBVCfT (ORCPT ); Tue, 21 Feb 2006 21:35:19 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750862AbWBVCfT (ORCPT ); Tue, 21 Feb 2006 21:35:19 -0500 Received: from ns1.siteground.net ([207.218.208.2]:27267 "EHLO serv01.siteground.net") by vger.kernel.org with ESMTP id S1750725AbWBVCfS (ORCPT ); Tue, 21 Feb 2006 21:35:18 -0500 Date: Tue, 21 Feb 2006 18:35:48 -0800 From: Ravikiran G Thirumalai To: Andrew Morton Cc: Nick Piggin , clameter@engr.sgi.com, linux-kernel@vger.kernel.org, shai@scalex86.org Subject: Re: [patch] Cache align futex hash buckets Message-ID: <20060222023548.GB3635@localhost.localdomain> References: <20060220233242.GC3594@localhost.localdomain> <43FA8938.70006@yahoo.com.au> <43FBB2E8.2020300@yahoo.com.au> <20060221180845.79a44449.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20060221180845.79a44449.akpm@osdl.org> User-Agent: Mutt/1.4.2.1i X-AntiAbuse: This header was added to track abuse, please include it with any abuse report X-AntiAbuse: Primary Hostname - serv01.siteground.net X-AntiAbuse: Original Domain - vger.kernel.org X-AntiAbuse: Originator/Caller UID/GID - [0 0] / [47 12] X-AntiAbuse: Sender Address Domain - scalex86.org X-Source: X-Source-Args: X-Source-Dir: Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1594 Lines: 42 On Tue, Feb 21, 2006 at 06:08:45PM -0800, Andrew Morton wrote: > Nick Piggin wrote: > > > > Christoph Lameter wrote: > > > On Tue, 21 Feb 2006, Nick Piggin wrote: > > > > > > > > >>Ravikiran G Thirumalai wrote: > > >> > > >>>Following change places each element of the futex_queues hashtable on a > > >>>different cacheline. Spinlocks of adjacent hash buckets lie on the same > > >>>cacheline otherwise. > > >>> > > >> > > >>It does not make sense to add swaths of unused memory into a hashtable for > > >>this purpose, does it? > > > > > > > > > It does if you essentially have a 4k cacheline (because you are doing NUMA > > > in software with multiple PCs....) and transferring control of that > > > cacheline is comparatively expensive. > > > > > > > Instead of 1MB hash with 256 entries in it covering 256 cachelines, you > > have a 1MB hash with 65536(ish) entries covering 256 cachelines. > > > > Good (if accidental point). Kiran, if you're going to gobble a megabyte, > you might as well use all of it and make the hashtable larger, rather than > just leaving 99% of that memory unused... Yes, good (intentional :) ) point. I am rerunning my tests with a larger hash slot. (As large as the padding takes away). If we get the same or better results, we can just increase the hash slots. Thanks, Kiran - 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/