From owner-freebsd-hackers Fri Jul 31 15:56:22 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA18011 for freebsd-hackers-outgoing; Fri, 31 Jul 1998 15:56:22 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from dingo.cdrom.com (dingo.cdrom.com [204.216.28.145]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA18006 for ; Fri, 31 Jul 1998 15:56:20 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost [127.0.0.1]) by dingo.cdrom.com (8.8.8/8.8.5) with ESMTP id PAA01058; Fri, 31 Jul 1998 15:54:44 -0700 (PDT) Message-Id: <199807312254.PAA01058@dingo.cdrom.com> X-Mailer: exmh version 2.0zeta 7/24/97 To: Craig Spannring cc: chanders@timing.com (Craig Anderson), freebsd-hackers@FreeBSD.ORG Subject: Re: Using FreeBSD for data acquisition? (long) In-reply-to: Your message of "Fri, 31 Jul 1998 13:59:56 PDT." <199807312059.NAA18156@bangkok.office.cdsnet.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 31 Jul 1998 15:54:44 -0700 From: Mike Smith Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > > That pretty much matches what I'm seeing for a server I'm writing. > My server must respond in less than 3 seconds. Normally it is able > to respond in a millisecond or two, but occasionally it takes around > 1000 milliseconds. > > I've duplicated the problem occurring even under real time priority > with no other processes on the machine except for init, pagedaemon, > vmdaemon, update, and one csh. I had virtual memory disabled to > eliminate any paging activity. If you're doing this over the net, it looks like you're seeing the TCP slow-start. Try turning it off; it makes a huge difference. 8) > I tracked the problem down to the fact that I'm not getting any CPU > time for a substantial (at least up to 500 milliseconds) amount of > time even though I'm using rtprio and there is nothing else that > should be running. So what is running? > In regards to your question of why 0.100 seconds is so common, FreeBSD > uses a 10msec quantum. I suspect if you change kern.quantum to some > other value then the other value will become more common. 0.1s != 10msec. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message