PCBUILD Archives

Personal Computer Hardware discussion List

PCBUILD@LISTSERV.ICORS.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"James E. Griffin" <[log in to unmask]>
Reply To:
PCBUILD - Personal Computer Hardware discussion List <[log in to unmask]>
Date:
Sun, 4 Apr 1999 18:21:34 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (59 lines)
Can it be done? Yes. Should you do it? Maybe. Normally, dual processor
servers are reserved for mission critical intensive computational
environments. In this environment, the cost of hardware is the LEAST of
your concerns. Do you have access to another dual processor server that
you can run a control test on? The acid test is if a known good dual
server running exactly the same problem returns the same answer as your
Celeron system, every time. Lots of bragging rights here if you do it
right.

If it's to learn on, it's worth your while; if for no other reason than
to find out what the limits of your abilities, test equipment and such
is. Taking a bunch of spare parts and making a network out of them --
eventually -- will teach you a great deal about electrical engineering,
providing you don't give up in frustration first.

If the intent was for anything like mission critical, and you're not
into EE, it would be a bad move. Could you look at someone else and ask
them to trust their needs to such a system? Depends on how critical
their stuff is.

How much EE background do you have? Given the gumption, knowledge and
equipment, you can interface just about anything with anything else.
You'll rapidly become familiar with the law of diminishing returns.

All this said, I'm a big fan of people learning by doing. Dual
processors take a bit of work, even with Intel approved CPU's. Onboard
SCSI compounds compatibility problems but gives you massive flexibility;
get real familiar with the NT hardware compatibility list.

From my point of view AT LEAST YOU'RE NOT ASKING TO OVERCLOCK THE
CELERONS YET. I've run into random strangeness trying to get dual's to
work at over 300 speed. Fortunately, the Asus is a real good board.
Running the Celerons at less than screaming maximum will give you some
reliability leeway.

Make sure ALL the hardware will work with NT; check the compatibility
lists from Microsoft and the various manufacturer's of the peripherals.
Patience, and a willingness to ask seemingly dumb questions will be your
biggest assets. The no-spin site has a bunch of useful stuff which you
should check out, on your project you'll get familiar with it and more.
Lastly, I'd recommend a continuous burn-in time of at least 72 hours.

All this may cost more money than you expect. Results could be a
learning experience second to none. Let the list know what's up.
>
> I want to put together a powerful but inexpensize server to run here at home
> with Windows NT server 4. I have the opportunity to buy a second-hand Asus
> dual processor LX board with Scsi on board. What I would like to know is:
> has anyone heard of or tried running dual celerons? I want a dedicated
> server for my cable internet connection, my home network, and possibly to
> run a file and print server, ftp server, and a gaming server
--
Prometheus was the first Beta Tester.
Sisyphus was the Corinthian Systems Administrator.

               PCBUILD mailing list is brought to you by:
                            The NOSPIN Group
                            http://nospin.com

ATOM RSS1 RSS2