From owner-freebsd-current@FreeBSD.ORG Tue Jun 1 19:46:18 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8FEB416A4CE; Tue, 1 Jun 2004 19:46:18 -0700 (PDT) Received: from sdf.lonestar.org (ol.freeshell.org [192.94.73.20]) by mx1.FreeBSD.org (Postfix) with ESMTP id 58D7843D49; Tue, 1 Jun 2004 19:46:18 -0700 (PDT) (envelope-from lukas@sdf.lonestar.org) Received: from sdf.lonestar.org (IDENT:lukas@sdf.lonestar.org [192.94.73.1]) by sdf.lonestar.org (8.12.10/8.12.10) with ESMTP id i522kCj6019046; Wed, 2 Jun 2004 02:46:12 GMT Received: (from lukas@localhost) by sdf.lonestar.org (8.12.10/8.12.8/Submit) id i522kCS2008812; Tue, 1 Jun 2004 19:46:12 -0700 (PDT) Date: Tue, 1 Jun 2004 19:46:12 -0700 (PDT) From: Luke X-X-Sender: lukas@sdf.lonestar.org To: Robert Watson In-Reply-To: Message-ID: References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed cc: freebsd-current@freebsd.org Subject: Re: Cap on network speed in CURRENT? X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: LukeD@pobox.com List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 02 Jun 2004 02:46:18 -0000 >> This is indeed a very slow host, by modern standards. It's a fussy old >> Pentium 166MHz that I've put through hell for the last eight years or >> so. >> >> I'm building a kernel with witness and invariants turned off. Building >> a kernel usually takes overnight. I'll try it out tomorrow and see how >> it goes. If those testing programs don't require X, I'll see about >> installing one of them to help me figure out what's going on. The iperf >> site looks like it's got some helpful information that I should read >> too. > > Also, try disabling SMP and possibly APIC support if it's a UP box. > There's a measurable performance overhead to compiling with SMP support, > since mutexes have to be compiled with locked instructions. This is a dual processor box, but I gave up and took SMP and APIC out of the kernel back in mid-April because the system would always panic on startup if I had both multiprocessor support and a RAID array defined at the same time. I made a thread about it on freebsd-questions but got no responses, so in complete bafflement I just turned off the multiprocessor support and rebuilt everything. The RAID was more important to me. http://docs.FreeBSD.org/cgi/mid.cgi?Pine.NEB.4.58.0404292248590.14141 I wouldn't be a bit surprised if that's a hardware problem. This old motherboard has a lot of quirks and I think it's probably going to die the next time I open the case. Thanks for the tip though. I wouldn't have thought of that.