From owner-freebsd-sparc64@FreeBSD.ORG Wed Sep 27 19:28:24 2006 Return-Path: X-Original-To: freebsd-sparc64@freebsd.org Delivered-To: freebsd-sparc64@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 93B5B16A412 for ; Wed, 27 Sep 2006 19:28:24 +0000 (UTC) (envelope-from jahnke@sonatabio.com) Received: from smtp.wizwire.com (smtp.wizwire.com [209.218.100.6]) by mx1.FreeBSD.org (Postfix) with ESMTP id AB9D843D58 for ; Wed, 27 Sep 2006 19:28:23 +0000 (GMT) (envelope-from jahnke@sonatabio.com) Received: from pinot.fmjassoc.com (209.218.101.53.bvi2.wizwire.com [209.218.101.53]) by smtp.wizwire.com (8.13.1/8.13.1) with ESMTP id k8RJRSMd006168 for ; Wed, 27 Sep 2006 12:27:29 -0700 From: Frank Jahnke To: freebsd-sparc64@freebsd.org Content-Type: text/plain Date: Wed, 27 Sep 2006 12:26:10 -0700 Message-Id: <1159385171.850.237.camel@pinot.fmjassoc.com> Mime-Version: 1.0 X-Mailer: Evolution 2.6.3 FreeBSD GNOME Team Port Content-Transfer-Encoding: 7bit X-WizWire-MailScanner-Information: Please contact the ISP for more information X-WizWire-MailScanner: Found to be clean X-MailScanner-From: jahnke@sonatabio.com Subject: Terrible hme throughput X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Sep 2006 19:28:24 -0000 I am seeing terrible network throughput on my Ultra 1E+ (a "fast" 1E clone). It uses the hme interface; throughputs using scp on a large file are about 1.3MB/s. Transferring the same file using scp between two different, faster computers gives me a throughput of about 8-9MB/s. This is a 100MB/s network, so that seems OK. ifconfig shows that hme is set properly (as far as I can tell): hme0: flags=8843 mtu 1500 options=b inet6 fe80::240:dcff:fe3b:21d8%hme0 prefixlen 64 scopeid 0x1 inet 192.168.0.103 netmask 0xffffff00 broadcast 192.168.0.255 ether 00:40:dc:3b:21:d8 media: Ethernet autoselect (100baseTX ) status: active and the MTUs are the same for all machines (and all are FreeBSD 6.1). Yes, I know that scp has overhead, but I find it hard to believe that the box could not keep up with it (it is a blazing 300MHz, after all), particularly for the throughput reduction I see. Any suggestions on how I can increase the network performance of my 1E+? Frank