Date: Wed, 15 Dec 1999 01:12:54 -0700 From: Brett Glass <brett@lariat.org> To: David Scheidt <dscheidt@enteract.com> Cc: Jamie Bowden <ragnar@sysabend.org>, Terry Lambert <tlambert@primenet.com>, noslenj@swbell.net, chat@FreeBSD.ORG Subject: Re: dual 400 -> dual 600 worth it? Message-ID: <4.2.0.58.19991215010917.048dfae0@localhost> In-Reply-To: <Pine.NEB.3.96.991214215754.44414A-100000@shell-1.enteract. com> References: <4.2.0.58.19991214174918.04736140@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
At 09:07 PM 12/14/1999 , David Scheidt wrote: > > Multiprocessing has always been a stopgap measure to get extra performance > > out of a machine until uniprocessors caught up. The diminishing returns > >But uniprocessors will never catch up. Actually, uniprocessors often do best in price/performance, because multiprocessor servers are priced so high and CPUs represent such a large precentage of the price of the system. >The glue needed to build an N-way >machine will always be less expensive than N uniprocessor boxes. Not so. The special chip sets are usually priced at a premium. > > make tightly coupled multiprocessing far less desirable than loosely > > coupled (or uncoupled!) distributed computing. > >For some applications loosely coupled multi-processing makes sense. For >others, like operations on one datastream, it doesn't. Actually, a Web page that draws images from several servers via IMG tags is very much like an "operation on one datastream," very neatly distributed. --Brett To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4.2.0.58.19991215010917.048dfae0>