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:
John Chin <[log in to unmask]>
Reply To:
PCBUILD - Personal Computer Hardware discussion List <[log in to unmask]>
Date:
Sat, 30 Jan 1999 02:42:26 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (58 lines)
>
>Just to add a bit to this, while it is true only one unmodified Celeron
>will work per system, dual Celerons can be used in a dual processor
>setup, but involves heavy modifications.  I'm not trying to be
>argumentative just wanted to add this little blurb lifted  from Tom's
>Hardware site. . . :
>
>"There is no reason for buying Pentium II anymore, you can even run the
>Socket370 Celerons as dual processors now! Plug them into a modified
>Slot1-converter card (like e.g. MSI’s MS-6905), as described by Tomohiro
>Kawada at http://kikumaru.w-w.ne.jp/pc/celeron/index_e.html. "
>
>Have I tried it?  Not in a million years!  This is way beyond me.  Just
>passing along info....



Jury-rigging a couple of Celerons for scalability in
a system with a critical mission suggests frivolity
or desperation. Currently, scant consumer software is
written for parallel processing. Unless the installation
is done by an electrical engineer or qualified technician
providing continuing support, I doubt many would entrust
their important work on such a system. A few hundred
dollars of savings is not worth the effort or risk.

Even in a less critical use, software is buggy enough
without introducing additional hardware variables.  I
always regret, after the fact, the time and anxiety wasted
trying to recover after "optimizing" a system just a
little too close to the hardware's limits. There's always
some Murphy's Law that will arise when tolerances are
minimized.

I admire people like Tom Pabst who push the envelope,
mostly for their hobbyist, empirical research.  They
are performing a public service as consumer advocates.
But, blindly implementing their discoveries without fully
understanding the process is foolhardy. In the real world,
lasting benefits are only realized when the concepts have
been validated, the engineering perfected and the product
(and price) is mainstreamed.

Perhaps the disinclination to tinker with a properly
engineered system comes with maturity, age, or the lack
of spare time.

Based upon my own experience, I would not rely on such
a system, except to play a game, having seen no actual
evidence or personal endorsement here to suggest otherwise.

Sincerely,

John Chin

        PCBUILD only works if you contribute. Send your messages
             to be posted to: [log in to unmask]

ATOM RSS1 RSS2