Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 2 Jul 2002 15:20:08 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 2 Jul 2002 15:20:07 -0400 Received: from h64-251-67-69.bigpipeinc.com ([64.251.67.69]:21772 "HELO kelownamail.packeteer.com") by vger.kernel.org with SMTP id ; Tue, 2 Jul 2002 15:20:06 -0400 From: "Stephane Charette" To: "linux-kernel@vger.kernel.org" Date: Tue, 02 Jul 2002 12:22:34 -0700 Reply-To: "Stephane Charette" X-Mailer: PMMail 2000 Standard (2.20.2502) For Windows 2000 (5.0.2195;2) MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Subject: performance impact of using noapic Message-Id: <20020702192006Z316884-686+254@vger.kernel.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1675 Lines: 34 I was reading through "FreeBSD Versus Linux Revisited" by Moshe Bar at "http://www.byte.com/documents/s=1794/byt20011107s0001/1112_moshe.html". One paragraph in particular caught my eye: On the Linux side, I attached all interrupts coming from the network adaptor to one CPU. With the new TCP/IP stack in the 2.4 kernels this really becomes necessary. Otherwise, you might find the incoming packets arranged out of order, because later interrupts are serviced (on another CPU) before earlier ones, thus requiring a reordering further down the handling layers. Is this a widely-known issue? Or is this simply theory? I'd never heard this mentionned until I read the article. I ran some web-based performance tests with the 2.4.19-pre9-SMP kernel on a dual-CPU Athlon 1600MHz box, and found that running with "noapic" actually improved network performance. (Negligable -- only 1% improvement in the small webstone-based test I ran.) As I wrote in another post concerning performance from earlier today, the actual values of my performance tests are not important -- the trend is what I wish to higlight. My questions are: 1) am I right in thinking that "noapic" will force all interrupts to be handled by 1 CPU? 2) how would you force all interrupts from only 1 hardware device (and not all devices) to be handled by 1 processor, as hinted in the paragraph quoted above? Thanks, Stephane Charette - 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/