From owner-freebsd-net Sun Oct 20 14:53:46 2002 Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 29EF237B401 for ; Sun, 20 Oct 2002 14:53:45 -0700 (PDT) Received: from mail.sandvine.com (sandvine.com [199.243.201.138]) by mx1.FreeBSD.org (Postfix) with ESMTP id 94EE743E91 for ; Sun, 20 Oct 2002 14:53:44 -0700 (PDT) (envelope-from don@sandvine.com) Received: by mail.sandvine.com with Internet Mail Service (5.5.2653.19) id <42S9VBZM>; Sun, 20 Oct 2002 17:53:43 -0400 Message-ID: From: Don Bowman To: 'Petri Helenius ' , Don Bowman , "'Kevin_Stevens@pursued-with.net '" Cc: "'freebsd-net@FreeBSD.ORG '" Subject: RE: ENOBUFS Date: Sun, 20 Oct 2002 17:53:41 -0400 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org From: Petri Helenius >> >> Clearly I will have some tuning ahead, and likely I will not = succeed, >> but for sure my 1U XEON with 6 gigabit nics will work very hard >> for its living :) >> >Which NICs seem to work best here? I=B4ve been playing more with em >and it seems that the time spent in interrupts is quite high, I=B4m = seeing >15-17% for 300Mbps on 2.4 Xeon. This number seems to stay the same >whether I=B4m running UP or SMP kernel with 4.7-STABLE. > >Does "giant" in 4.X SMP context mean that the other CPU is idling = while >the other is either servicing interrupts or running kernel code? > >What would be the best course of action to implement optimizations >possible with later chips like 82546 to the em driver? Talk to Intel? Well, I'm definitely finding that I have more CPU free when using the broadcom BCM570X NIC (bge) than the Intel 8254X NIC (em). http://www.etestinglabs.com/main/reports/broadcom.asp http://www.etestinglabs.com/main/reports/3com.asp has a benchmark of the broadcom versus the intel. This is on a win2k platform. These benchmarks were paid for by broadcom,=20 so take them with a grain of salt. As for tuning the driver for either chip, you will need an NDA to get the documentation. --don (don@sandvine.com www.sandvine.com) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message