Date: Thu, 21 Mar 2002 13:07:10 +1030 From: Greg 'groggy' Lehey <grog@FreeBSD.org> To: Patrick Thomas <root@utility.clubscholarship.com> Cc: questions@freebsd.org Subject: Re: is there overhead in SMP machines related to SMP ? Message-ID: <20020321130710.B49090@wantadilla.lemis.com> In-Reply-To: <20020320083947.M8701-100000@utility.clubscholarship.com> References: <20020320083947.M8701-100000@utility.clubscholarship.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[Format recovered--see http://www.lemis.com/email/email-format.html] Output wrapped. On Wednesday, 20 March 2002 at 8:45:27 -0800, Patrick Thomas wrote: > > I have a quad-zeon 500. I loaded 4.5-RELEASE on it and there are perhaps > 80-90 processes now running. `vmstat` output looks like: > > procs memory page disks faults cpu > r b w avm fre flt re pi po fr sr aa0 md0 in sy cs us sy > id > 1 0 0 29056 329772 25 0 0 0 32 4 0 0 235 145 11 0 0 100 > 0 0 0 29056 329772 4 0 0 0 0 0 0 0 282 21 8 0 4 96 > 0 0 0 29056 329772 5 0 0 0 0 0 0 0 260 30 12 0 3 97 > 0 0 0 29056 329772 5 0 0 0 0 0 0 0 287 66 31 0 3 97 > 0 0 0 29056 329756 7 0 0 0 0 0 0 0 290 85 24 0 4 96 > 0 0 0 29056 329740 3 0 0 0 0 0 1 0 276 58 21 0 3 97 > 0 0 0 29056 329724 3 0 0 0 0 0 0 0 293 52 21 0 4 96 > 0 0 0 29056 329724 7 0 0 0 0 0 0 0 281 104 28 0 3 97 > > Note the 96-96% idle cpu. This seems very odd, since my laptop, and other > single processor 4-500 mhz (pIII) systems are running many more processes, > but their idle is between 98-99. > > However, when my laptop goes to 95-96, it is because "user" time goes up, > but on my quad system it is hovering at 96-97 because "system" time is up. > > So should I assume that this is just an artifact of using SMP, and that > whereas my laptop starts at 99-100 idle and goes down from there, this > will just start at 96-97 idle and go down from there (but taking many more > processes to drop me from 96 to 95 than it takes to drop my laptop from 99 > to 98) ? > > Or is something else going on ? Probably something else is going on. You can't really tell just from vmstat. Note also that it's not very interesting to optimize the idle state. 4.5 has a number of inefficiencies in SMP; we're addressing some of them in 5.0. Greg -- When replying to this message, please copy the original recipients. If you don't, I may ignore the reply. For more information, see http://www.lemis.com/questions.html See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020321130710.B49090>