From owner-freebsd-hardware Thu Feb 25 18:15: 7 1999 Delivered-To: freebsd-hardware@freebsd.org Received: from pluto.plutotech.com (mail.plutotech.com [206.168.67.137]) by hub.freebsd.org (Postfix) with ESMTP id 855C214DFA for ; Thu, 25 Feb 1999 18:15:05 -0800 (PST) (envelope-from drew@pluto.plutotech.com) Received: (from drew@localhost) by pluto.plutotech.com (8.9.1/8.9.1) id TAA12214; Thu, 25 Feb 1999 19:14:49 -0700 (MST) (envelope-from drew) Date: Thu, 25 Feb 1999 19:14:49 -0700 (MST) From: Drew Eckhardt Message-Id: <199902260214.TAA12214@pluto.plutotech.com> To: hardware@freebsd.org Subject: Re: Intel vs. 3com vs DEC chip network cards X-Newsgroups: pluto.freebsd.hardware In-Reply-To: <3.0.32.19990223163334.00692d68@advantinc.com> Organization: Cc: Sender: owner-freebsd-hardware@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In article <3.0.32.19990223163334.00692d68@advantinc.com> you write: >Are there any reliability or performance issues on using one type over the >other? Some DEC chip + PHY combinations bogusly detect collisions when run in full-duplex 100baseT mode. The driver drops these packets on the floor, the TCP code gets no ACK, and eventually things are retransmitted resulting in "full duplex (autodetected or manually specified)" operation which has best-case performance that's two orders of magnitude worse than "half duplex". The Intel boards work well (read as both robust and fast), don't cost much (I paid $45 a pop for my pair of 10/100 boards without wake-on-lan), and get my vote. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message