From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753877AbYLOODR (ORCPT ); Mon, 15 Dec 2008 09:03:17 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753187AbYLOODE (ORCPT ); Mon, 15 Dec 2008 09:03:04 -0500 Received: from nlpi053.sbcis.sbc.com ([207.115.36.82]:49266 "EHLO nlpi053.prodigy.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753160AbYLOODD (ORCPT ); Mon, 15 Dec 2008 09:03:03 -0500 Date: Mon, 15 Dec 2008 08:02:47 -0600 (CST) From: Christoph Lameter X-X-Sender: cl@quilx.com To: Nick Piggin cc: Linux Kernel Mailing List , Linux Memory Management List , bcrl@kvack.org, list-linux-mm@kvack.org Subject: Re: [rfc][patch] SLQB slab allocator In-Reply-To: <20081214230407.GB7318@wotan.suse.de> Message-ID: References: <20081212002518.GH8294@wotan.suse.de> <20081214230407.GB7318@wotan.suse.de> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Score: -2.6 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 15 Dec 2008, Nick Piggin wrote: > > Does this mean that SLQB is less efficient than SLUB for off node > > allocations? SLUB can do off node allocations from the per cpu objects. It > > does not need to make the distinction for allocation. > > I haven't measured them, but that could be the case. However I haven't > found a workload that does a lot of off-node allocations (short lived > allocations are better on-node, and long lived ones are not going to > be so numerous). A memoryless node is a case where all allocations will be like that. > That's more complexity, though. Given that objects are often hot when > they are freed, and need to be touched after they are allocated anyway, > the simple queue seems to be reasonable. Yup. > This case does improve the database score by around 1.5-2%, yes. I > don't know what you mean exactly, though. What case, and what do you > mean by bad cache unfriendly programming? I would be very interested > in improving that benchmark of course, but I don't know what you > suggest by keeping cachelines hot in the right way? What I was told about the database test is that it collects lists of objects from various processors that are then freed on a different processor. This means all objects are cache cold.