From owner-freebsd-net Fri Oct 18 14:32: 2 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 837C837B401 for ; Fri, 18 Oct 2002 14:32:01 -0700 (PDT) Received: from overlord.e-gerbil.net (e-gerbil.net [64.186.142.66]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2635443E88 for ; Fri, 18 Oct 2002 14:32:01 -0700 (PDT) (envelope-from ras@e-gerbil.net) Received: by overlord.e-gerbil.net (Postfix, from userid 1000) id 961FA15E4B; Fri, 18 Oct 2002 17:31:54 -0400 (EDT) Date: Fri, 18 Oct 2002 17:31:54 -0400 From: Richard A Steenbergen To: Kevin Stevens Cc: Don Bowman , freebsd-net@FreeBSD.ORG Subject: Re: ENOBUFS Message-ID: <20021018213154.GO26000@overlord.e-gerbil.net> References: <20021018135434.Y40012-100000@babelfish.pursued-with.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20021018135434.Y40012-100000@babelfish.pursued-with.net> User-Agent: Mutt/1.3.27i 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 On Fri, Oct 18, 2002 at 02:04:44PM -0700, Kevin Stevens wrote: > > First, you're only pushing packets, so you are only talking a potential > 1GB, not two. > > Second, sending minimum-size packets, while a best-case metric for pps, is > a worst-case metric for throughput. I don't think that you can conclude > that 20% theoretical bandwidth utilization at minimum packet size is poor > performance; in fact it seems pretty good to me. Extrapolating from those > numbers, if the packets were five times larger (320b), you'd hit > theoretical maximum throughput. Obviously that won't happen, your pps > numbers will go down as the packet size goes up, but it does indicate you > have some headroom, even without going to jumbo frames if the card pushed > five times fewer pps at 1500 byte frames you'd max out the throughput. > > You can't tell very much from a single data point like that, but what you > can infer doesn't seem to me to be bad at all. Counting overhead for the preamble, SFD, the ethernet headers, the checksum, the interframe gap, and the padding required to produce a minimum sized ethernet frame, the equivalent of no less than 84 bytes is transmitted for any frame. That makes the maximum unidirectional packets/sec on GigE 1.488Mpps. 400kpps is a little under 27% of this. In "normal" use that's probably good enough, but it's hardly what I'd call line rate. :) -- Richard A Steenbergen http://www.e-gerbil.net/ras PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message