Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 23 May 1998 23:38:10 -0400
From:      kriston@ibm.net (Kriston J. Rehberg)
To:        freebsd-advocacy@FreeBSD.ORG
Subject:   Re: FreeBSD A Solution For Business
Message-ID:  <7442-Sat23May1998233810-0400-kriston@ibm.net>
In-Reply-To: <199805230527.WAA03120@osprey.grizzly.com>
References:  <01bd85e0$2dccb1c0$f820aace@eliot.pacbell.net> <199805230337.UAA02883@osprey.grizzly.com> <3626-Sat23May1998005302-0400-kriston@ibm.net> <199805230527.WAA03120@osprey.grizzly.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Mark Diekhans writes:

>The rational here is that user-level threads eliminate much of the process
>context switching overhead.  Not having seen or done performance measurements,
>I can't say if its significant to a http server or not, I discussed it with
>people in charge of selecting systems who felt a threaded server was required
>(but then again, they didn't have any measurements to back it up).  Personally,
>I would just add another pentium running apache if the first one got bogged
>down.

Yeah, but I thought the problem is that you don't have a guarantee
that the scheduler will run your Apache process on the second Pentium.
I don't tend to believe that merely adding processors automatically
makes everything faster.  Sure, if your system starts and stops lots
of processes, you will get an advantage.  But if it's a persistent
process like a web server, is there really a guarantee that it will go
onto another processor on your system?

But as for the user-level threads in FreeBSD, can they be spread out
over >1 processor, or does that require kernel threads?

Kris

-- 
Kriston J. Rehberg
AOL: Kriston                        http://kriston.net/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-advocacy" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7442-Sat23May1998233810-0400-kriston>