From owner-freebsd-stable Sun Feb 21 19:10:22 1999 Delivered-To: freebsd-stable@freebsd.org Received: from freya.circle.net (freya.circle.net [209.95.95.2]) by hub.freebsd.org (Postfix) with ESMTP id 9780011EAC for ; Sun, 21 Feb 1999 19:10:20 -0800 (PST) (envelope-from tcobb@staff.circle.net) Received: by freya.circle.net with Internet Mail Service (5.5.2232.9) id ; Sun, 21 Feb 1999 22:09:33 -0500 Message-ID: From: tcobb@staff.circle.net To: doconnor@gsoft.com.au Cc: stable@FreeBSD.ORG Subject: RE: SMP and libc_r Date: Sun, 21 Feb 1999 22:09:32 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2232.9) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > On 22-Feb-99 tcobb@staff.circle.net wrote: > > Should I expect to see performance improvements with SMP > > and a multithreaded app using libc_r? The app in question > No :( > Not yet anyway :) That's what I was afraid of. This awaits kernel-threads, eh? > > is MySQL and doing the sql benchmarks with an SMP (dual pII 350) > > and non-smp kernel for the same box, I see much faster > > times with the non-smp version. > The SMP kernel spends more time in system locks over the > entire kernel which is a bit > annoying. SMP works much better for things like multiple > CPU intensive processes, it > works really badly on multiple IO intensive apps (because > the all wait for the big kernel > lock) I find it strange that it should perform worse than a single CPU kernel on the same tests, though. Correct me if I'm wrong, but, if I throw alot of memory at MySQL and get it to cache the more frequently-accessed tables then I should be able to work around this? -Troy Cobb Circle Net, Inc. http://www.circle.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message