From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [PATCH net-next] net: allocate skbs on local node Date: Thu, 14 Oct 2010 12:27:52 -0700 Message-ID: <20101014122752.21dd4eaf.akpm@linux-foundation.org> References: <1286838210.30423.128.camel@edumazet-laptop> <1286839363.30423.130.camel@edumazet-laptop> <1286859925.30423.184.camel@edumazet-laptop> <20101011230322.f0f6dd47.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: Eric Dumazet , David Miller , netdev , Michael Chan , Eilon Greenstein , Christoph Hellwig , Christoph Lameter To: Tom Herbert Return-path: Received: from smtp1.linux-foundation.org ([140.211.169.13]:35482 "EHLO smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753146Ab0JNT2h (ORCPT ); Thu, 14 Oct 2010 15:28:37 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Thu, 14 Oct 2010 08:31:01 -0700 Tom Herbert wrote: > > This is all conspicuously hand-wavy and unquantified. __(IOW: prove it!) > > > > The mooted effects should be tested for on both slab and slub, I > > suggest. __They're pretty different beasts. > > -- > > Some results running netper TCP_RR test with 200 instances, 1 byte > request and response on 16 core AMD using bnx2x with one 16 queues, > one for each CPU. > > SLAB > > Without patch 553570 tps at 86% CPU > With patch 791883 tps at 93% CPU > > SLUB > > Without patch 704879 tps at 95% CPU > With patch 775278 tps at 92% CPU > > I believe both show good benfits with patch, and it actually looks > like the impact is more pronounced for SLAB. I would also note, that > we have actually already internally patched __netdev_alloc_skb to do > local node allocation which we have been running in production for > quite some time. > Yes, that's a solid gain. Can we think of any hardware configuration for which the change would be harmful? Something with really expensive cross-node DMA maybe?