Re: Uhhh.... Re: Pentium II processor..

[Christopher Barry <cbarry@2xtreme.net>]

If _BOTH_ of the processors just up and died like that, chances are:

a) You had the voltage improperly set.
b) Did you buy OEM? You may have bought 333s, overclocked to 400 and
remarked.
c) You damaged some of the voltage and/or other control circuitry via
static discharge. A 200V discharge can damage and/or kill a great many
semiconductors. Have you ever touched a doorknob and been shocked?
That's typically 9000V. You could have given several parts of the
motherboard and/or processors some lovely 200V-1000V shocks by
installing them with your bare hands and without a wrist grounding
strap, and never felt a thing.

Jeremy Tregunna wrote:
> 
> > Geeees,
> >
> > Ummm... look. The Pentium II and Celeron 266 - 300 are the exact same
> > chip, just the card that the Pentium II is on has two 256k cache SRAMs
> > on it, operating at half processor frequency. They both have MMX, blah,
> > blah, etc. They are the same exact chip, made on the same assembly line,
> > just one is mounted to a little card with some cache on it, and the
> > other one (Celeron) isn't. Now the Celeron 300A and 333 are slightly
> > different in that they have 128k built into the chip, which runs at full
> > speed.
> >
> > Now any Pentium II outperforms the Pentium Pro 200 at every single
> > conceivable test. There is nothing a Pro 200 does better than a Pentium
> 
> I've tested two P2 400's as my main network server at my house (which is
> the core of the county-wide extranet) and the first one couldn't handle
> the prime time hours, as soon as i hooked it up, traffic started to go
> through, etc, etc, and the P2 blew up in 4 minutes and 30 some odd
> seconds. The second one, however, lasted 16 hours (put it up at 11pm, and
> she blew up later on the next day). (Needless to say, the PPro was
> sluggish, big time, but it didn't blow up. And before you say it was a
> cooling problem, it wasn't, just trust me on that. Anyway, that spot on
> the network is not served by a cluster of two 21164 DEC Alpha's.)
> 
> > II. I don't know where you are getting your information, but I'd love to
> > see some web links or references or something.
> 
> Personal tests, I could put them up on the web, but then it would be just
> like talking over this mailing list. My tests, against yours. My machines
> have to withstand a hellova lot of shit, and for that I've found the RiSC
> based CPU to be the best (but can't always fit into the average person's
> budget).
> 
> >
> > Buying a Pentium II was not a mistake. This processor offers the most
> 
> The early Pentium 2's are well documented on having problems with them,
> this is not the case in newer P2's,

The PII 233-300 (.35 CMOS MICRON, klamath core) sucked a lot of power
and had heat problems. The 333-450 (.25 CMOS MICRON, deschutes core) are
low power and run fairly cool, and in no way could be considered any
more problematic than an Alpha or whatever else.

 but MMX has been a severe problem in
> the past (this is not limiting the problem to a P2 only, the Pentium w/MMX
> has had the same sort of problems, due to MMX).

And this is from your personal experience? I've never heard about the
presence of an MMX unit causing severe problems, or the MMX unit itself
containing serious errata.

> 
> > floating point performance of any processor for the money right now,
> > except for the Celeron(which as I just explained is a cheaper version of
> > the same thing), and depending on the benchmark may or may not offer
> > better price/performance over the K6. The fastest Pentium II at 450MHz
> > naturally kicks the crap out of any other CPU out there with the
> > exception of the Alpha and the latest UltraSPARC(depending on
> > benchmark).
> >
> 
> Celeron CPU's are in my opinion, better server CPU's than the P2
> (excluding the new, non-faulty Xeons which I have yet to test thouroully).
> But I have tested Celeron 300a against a P2 400, and the results just
> showed that the P2 was a bit more sluggish than the Celeron on certain
> things, yet outperformed Celeron (a few cases miserably) in others.

Check out spec.org and some popular server benchmark sites. The fact
that the PII has 4 times the cache of the newest Celerons is very
important for servers that have high throughput, and large ammounts of
main memory (256MB and up). As a matter of fact, the new Celerons seem
to perform on par with PIIs of the same speed on most nearly anything
(3D gaming, office app performance, etc...) _except_ for servers. If
this has been your experience, then well, that's all fine and fair, but
you may have misconfigured something.

> 
> This is just what I've found over my testing, if you don't like what I
> have to say, then dont' respond, PERSONAL TESTS.

Understood.
 
> --Jeremy Tregunna
> intrigue@golden.net +1 519 338-3591