From owner-freebsd-current@FreeBSD.ORG Sat May 14 19:20:19 2005 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 E81FB16A4CE; Sat, 14 May 2005 19:20:18 +0000 (GMT) Received: from lexi.siliconlandmark.com (lexi.siliconlandmark.com [209.69.98.4]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5EF8843D1F; Sat, 14 May 2005 19:20:18 +0000 (GMT) (envelope-from andy@siliconlandmark.com) Received: from lexi.siliconlandmark.com (localhost [127.0.0.1]) j4EJKFqO003905; Sat, 14 May 2005 15:20:15 -0400 (EDT) (envelope-from andy@siliconlandmark.com) Received: from localhost (andy@localhost)j4EJKB83003902; Sat, 14 May 2005 15:20:15 -0400 (EDT) (envelope-from andy@siliconlandmark.com) X-Authentication-Warning: lexi.siliconlandmark.com: andy owned process doing -bs Date: Sat, 14 May 2005 15:20:11 -0400 (EDT) From: Andre Guibert de Bruet To: Mike Jakubik In-Reply-To: <4603.172.16.0.199.1116094186.squirrel@172.16.0.1> Message-ID: <20050514150435.I3060@lexi.siliconlandmark.com> References: <20050505133250.GA73885@us.svf.stuba.sk> <20050507043712.GB28373@xor.obsecurity.org> <1318.172.16.0.199.1115846681.squirrel@172.16.0.1> <1424.172.16.0.199.1115851128.squirrel@172.16.0.1> <3973.172.16.0.199.1116028322.squirrel@172.16.0.1> <4603.172.16.0.199.1116094186.squirrel@172.16.0.1> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-Information: Please contact the ISP for more information X-SL-MailScanner: Found to be clean X-SL-SpamCheck: not spam, SpamAssassin (score=-2.548, required 6, autolearn=not spam, AWL 0.05, BAYES_00 -2.60) X-MailScanner-From: andy@siliconlandmark.com cc: freebsd-net@freebsd.org cc: Andre Oppermann cc: current@freebsd.org Subject: Re: Outgoing speed problems in -CURRENT (was: Re: SOLVED: Degraded TCP X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 14 May 2005 19:20:19 -0000 On Sat, 14 May 2005, Mike Jakubik wrote: > On Sat, May 14, 2005 12:49 am, Andre Oppermann said: >> Mike Jakubik wrote: > >>> I just tried your patch, but unfortunately it does not seem to affect >>> my problem, the speeds are identical. I am still getting half the >>> performance when sending out. The fact that this also occurs using >>> another machine and a different card, leads me to belive something is >>> broken in -CURRENT. >> >> net.inet.tcp.sendspace=65536 and try again. > > Im sorry, i mean to say receiving, not sending. I did try adjusting this, > but it made no difference. Just to recap: > > netio server running windows xp, connecting from fbsd: > > TCP/IP connection established. > Packet size 1 KByte: 41103 KByte/s > Packet size 2 KByte: 39601 KByte/s > Packet size 4 KByte: 43463 KByte/s > Packet size 8 KByte: 43734 KByte/s > Packet size 16 KByte: 43719 KByte/s > Packet size 32 KByte: 43656 KByte/s > > netio server running on fbsd, connecting from windows xp: > > TCP/IP connection established. > Packet size 1 KByte: 19347 KByte/s > Packet size 2 KByte: 21056 KByte/s > Packet size 4 KByte: 22845 KByte/s > Packet size 8 KByte: 23445 KByte/s > Packet size 16 KByte: 25082 KByte/s > Packet size 32 KByte: 25294 KByte/s > > So as we can see, i am getting nearly half the performance when sending to > the fbsd server. I get simillar results when substituting the windows > server with a fbsd 5.4 server, and using 100mbit cards on both. You surely must have meant 1000 Mbit cards. These numbers exceed the theoretical throughput of 100 Mbit networking (12500 KBytes/s). Do the numbers differ significantly if you try this test between your CURRENT machine and say, a FreeBSD 4.x or Linux 2.6 machine? Andy /* Andre Guibert de Bruet * 6f43 6564 7020 656f 2e74 4220 7469 6a20 */ /* Code poet / Sysadmin * 636f 656b 2e79 5320 7379 6461 696d 2e6e */ /* GSM: +1 734 846 8758 * 5520 494e 2058 6c73 7565 6874 002e 0000 */ /* WWW: siliconlandmark.com * Tormenting bytes since 1980. */