Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!spool.mu.edu!howland.reston.ans.net!newsfeed.internetmci.com!news.mathworks.com!nntp.primenet.com!uunet!inXS.uu.net!news.artisoft.com!usenet
From: Terry Lambert <terry@lambert.org>
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: dual processors
Date: Wed, 17 Jul 1996 19:16:56 -0700
Organization: Me
Lines: 30
Message-ID: <31ED9E98.61FB29C1@lambert.org>
References: <31C67EA0.9E0C259@unix.sri.com> <Pine.ULT.3.91.960705140128.3415E-100000@blaze.trentu.ca> <31DD7D6D.5A46CCCD@lambert.org> <4s3ok2$go3@news.abs.net> <31E5CE8D.4CDCA958@lambert.org> <4sgh8u$lc@anorak.coverform.lan>
NNTP-Posting-Host: hecate.artisoft.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 2.01 (X11; I; Linux 1.1.76 i486)
Brian Somers wrote:
] Terry Lambert (terry@lambert.org) wrote:
] : FreeBSD has supported low grain (kernel entrancy lock) SMP since
] : October 1994.
]
] So how are things these days ? Is the plan to go fully-preemptive
] - maybe with a run-time-configurable non-preemptive option to dodge
] any locking overheads on a uniprocessor machine ?
]
] It would be nice if you could summarize the position. Thanks.
I am personally avoiding the scheduler and the system startup/pmap
code and concentrating on kernel preemption, which I intend to
use in the UP case on my own machines, even if it never goes
mainstream. Historically, kernel preemption added a 160% performance
improvement on UFS in UnixWare 2.x in the UP case, even after lock
overhead was subtracted out. I believe it is a win, and necessary
for RT and kernel threading, in any event.
As far as direction, you should contact Peter Wemm and/or
Poul-Henning Kamp for details on what will and won't be
integrated into the mainstream code base, and when.
Regards,
Terry Lambert
terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.