From owner-freebsd-net Fri Jan 7 17: 8:18 2000 Delivered-To: freebsd-net@freebsd.org Received: from cs.rice.edu (cs.rice.edu [128.42.1.30]) by hub.freebsd.org (Postfix) with ESMTP id 7A03315854 for ; Fri, 7 Jan 2000 17:08:06 -0800 (PST) (envelope-from aron@cs.rice.edu) Received: (from aron@localhost) by cs.rice.edu (8.9.0/8.9.0) id TAA06197; Fri, 7 Jan 2000 19:08:03 -0600 (CST) From: Mohit Aron Message-Id: <200001080108.TAA06197@cs.rice.edu> Subject: Re: performance of FreeBSD-current as SMP To: weyrich@goodnet.com (Weyrich Computing Consulting) Date: Fri, 7 Jan 2000 19:08:03 -0600 (CST) Cc: freebsd-net@freebsd.org In-Reply-To: from "Weyrich Computing Consulting" at Jan 7, 2000 06:04:27 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > To hazard a guess, there is certain overhead in SMP. If your processing > mix is able to take advantage of concurrent processing, then you will see > a net performance boost, but not the expected doubling (due to the > overhead penalty). > Like I said, I'm just using one processor. So concurrent processing issues don't arise. What's surprising is that just configuring it as an SMP makes the performance go down by 22%. > The question is, does your benchmark allow concurrent > processing? If you have a single Network Interface Card, does it act as a > non-sharable resource that effectively precludes concurrent processing by > two web server processes? Your web server does spawn a new process for > each client session, and you are generating multiple client sessions > concurrently, right? > No, the webserver is event driven - does everything withing a singe process. - Mohit To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message