From owner-freebsd-current Fri Nov 6 15:15:06 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA19206 for freebsd-current-outgoing; Fri, 6 Nov 1998 15:15:06 -0800 (PST) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from janus.syracuse.net (janus.syracuse.net [205.232.47.15]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA19191 for ; Fri, 6 Nov 1998 15:14:52 -0800 (PST) (envelope-from green@unixhelp.org) Received: from localhost (green@localhost) by janus.syracuse.net (8.8.8/8.8.7) with ESMTP id SAA06426; Fri, 6 Nov 1998 18:14:17 -0500 (EST) Date: Fri, 6 Nov 1998 18:14:16 -0500 (EST) From: Brian Feldman X-Sender: green@janus.syracuse.net To: Alfred Perlstein cc: Cory Kempf , freebsd-current@FreeBSD.ORG Subject: Re: bad time slicing? Priorities? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Of course, noone would expect the kernel to spend all of that time in the scheduler, so for it to be real-time, people would be wondering why their CPU was at 40% when idle. Brian On Fri, 6 Nov 1998, Alfred Perlstein wrote: > > On Fri, 6 Nov 1998, Brian Feldman wrote: > > > You don't really understand the scheduling. Nice the xgalaga to 0 or -1, > > and try again. Plus, rc5des is running on BOTH CPU's (FreeBSD splits it of > > course and switches them around to have the best performance), FreeBSD is > > not "magic". > > > > Brian Feldman > > More like the fact that for every screen update you have a context switch > between the game and the X server several times a second, with 2 CPU > hungry monsters in the background they are bound to steal cycles. The > granularity of the scheduler is for interactive typing (ie. at a terminal > you wouldn't notice the 2 rc5's), not hi-rez/fast context switching gfx > games. It's something expected afaik. > > -Alfred > > > > > > On 6 Nov 1998, Cory Kempf wrote: > > > > > I have a system running 3.0 SMP, with 2 333MHz PII's. > > > > > > On this system, I am running two copies of the Bovine RSA client (rc5des) > > > (essentially two endless CPU bound tasks, niced down to 19) > > > > > > >From top: > > > PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU COMMAND > > > 23593 root 105 19 832K 344K RUN 0 23.4H 98.15% 98.15% rc5des > > > 23726 root 105 19 832K 336K CPU1 0 22.7H 97.89% 97.89% rc5des > > > > > > We are not swapping, or anything else obvious. > > > > > > If I run xgalaga (a game), which is being run at nice=5 for some > > > reason (not sure why, haven't looked into it), It doesn't seem to be > > > getting enough CPU time. > > > > > > Play is jerky and slow. > > > > > > This isn't what I expected. Expecially with two CPUs to play with. > > > The rc5des programs, should essentially not be running if higher > > > priority things are waiting to run, right? Certainly with two CPUs, I > > > would expect that the game would get time pretty much as soon as it > > > was ready to run, while the other tasks would fight over what was > > > left. > > > > > > So, do I just not understand how BSD does its scheduling? Or is there > > > actually something wrong? > > > > > > Thanks, > > > > > > +C > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message