From owner-cvs-all Thu Nov 5 20:15:22 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA14368 for cvs-all-outgoing; Thu, 5 Nov 1998 20:15:22 -0800 (PST) (envelope-from owner-cvs-all@FreeBSD.ORG) Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA14363; Thu, 5 Nov 1998 20:15:19 -0800 (PST) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.1/8.9.1) id UAA00826; Thu, 5 Nov 1998 20:15:03 -0800 (PST) (envelope-from dillon) Date: Thu, 5 Nov 1998 20:15:03 -0800 (PST) From: Matthew Dillon Message-Id: <199811060415.UAA00826@apollo.backplane.com> To: Marc Slemko Cc: "Dag-Erling C. =?iso-8859-1?Q?Sm=F8rgrav?=" , David Greenman , cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/conf options param.c src/sys/kern init_sysent.c syscalls.c syscalls.master uipc_syscalls.c src/sys/sys mbuf.h socket.h socketvar.h syscall-hide.h syscall.h syscall.mk sysproto.h src/sys/vm vm_object.c References: Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk :When I have access to a system I also plan to implement it in Apache to :see what sort of difference it makes. : :While the ideal implementation won't be until Apache 2.0, it will be :interesting to see just what impact this has with the current Apache. I'm sure it will make a huge difference in the benchmarks, but I doubt you would notice much of a difference in real-life performance except in the most extreme installations. Cpu's are cheap, physical I/O limits are usually more important because its the first thing web servers tend to hit. The only time cpu becomes an issue is when the dataset being served is small enough to fit in system memory. Then disk I/O becomes irrelevant and the issue becomes a pure cpu and network problem. Unfortunately, it is still mostly irrelevant because the cost of pushing a megabyte a second out to the internet is several thousand dollars a month and the cost of the cpu required to back that push is less then $200. The only situation that I can think of where sendfile() would be significant in a real-life situation is when it is used to serve an internal network... for example, to serve video or music mpegs off of a cheap machine inside an office. In this case your major cost is the machine and not the network, so being able to use a cheap cpu can save you significant money. -Matt Matthew Dillon Engineering, HiWay Technologies, Inc. & BEST Internet Communications & God knows what else. (Please include original email in any response) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message