Hardware Reviews

Jason Gurtz jason at tommyk.com
Fri Aug 27 09:40:47 PDT 2004

On 8/27/2004 06:17, Ian Molton wrote:

> ARGH not that horribly biased article *again*. THG is *crapola*

Maybe so, maybe not...  My take on THG is that it is OK for *very* basic
info in an "executive summery" type of way.  *All* reviews, *anywhere*
are biased because all people have opinion and no matter what the best
intentions may be, invariably end up corrupting even the best reviewer.
 There are certainly worse places to go then toms.  Maybe boy wonder
Anand "who's buying him today" La Shimpi is better.... hrm

> Intel CPU overheat protection is all on-chip. the AMD version requires
> co-operation from the mainboard. All current AMD boards support this.
> 1) INTEL  *paid* for that test
> 2) It used a NON temperature controlled motherboard even though
> controlled ones were (just) available at the time.

That word is key.  The article in question is very old now but *was*
relevant at that time because the AMD platform consisted mostly of
shitty VIA chipset boards that did not support the feature.  Only the
rich could afford to get the rare boards that had the actual AMD chipset
which was much better.  Nforce simply did not exist, were there even
rumors then?

All this, of course, has now changed and *much* higher quality AMD
platform boards are now generally available.

I would love to find a review site that has reviews that are *not* paid
for by someone or other.  There are a very few (Dan's data for one) that
are generally very high quality.  But even there IIRC, hardware is often
donated which means the review was paid for in a sense.

Sort of like Windows vs. Linux.  Invariably you get either a windows
loon who knows very little about Linux or a Linux nut who doesn't really
understand Windows (Open magazine is actually pretty good here).  Either
way, all the reader gets is a shitty waste of time.   ...unless it's
taken in context

If you know of good non-partisan hardware review sites I would love it
if you would please post the list.  I would honestly be very grateful
because, as it is now, the evaluation process requires reading multiple
reviews (none of which I trust) in order to ascertain the true meaning
and then, in addition, reading Mfr. spec sheets (which often omit
certain key details!) and then support documents to try and get an idea
of the scope and history of fixes for the product.  Also, if you have
definitive evidence that certain review are "paid for" by certain
companies I would be grateful for that too.  I could really use such
evidence in critical papers I write for school.

> 3) Heatsinks *dont* fall off, unless you violate the specs of the mounts
> on the board (as many heatsinks did at that time).

Except sometimes in shipping...  More commonly, the fan will fail which,
although not quite as catastrophic, can still result in problems.

> I *KNOW* its crap because I *have* put a heatsink on an AMD chip upside
> down (leaving a clear air gap between CPU and 'sink.). The board powered
> up, and after about 1-2 seconds powered down blinking a red 'CPU
> overheat' LED at me. (Nforce 2 board in shuttle SN45G) (thats the most
> recent, I've done it before on older boards).

Like I said, the quality of AMD platform boards (such as yours) has
improved dramatically since the time of that article which relegates the
video to "party stunt" status.  But, this is nearly 3 years after that
video was published.  It was not a party joke then, regardless who may
have paid for it.

> so, OK AMD systems shut down, but thats not really any worse than a
> system that runs uselessly slowly anyway.

Hmm, that really depends on the situation.  Just this last week the cpu
fan died on the mail server here at work (intel P-200MMX)  Actually, it
probably died about 2 or 3 weeks ago and the only reason I found out was
because of an intermittent beep that I couldn't quite locate in the
server room and user complaints of the web mail being "pretty slow".  I
use POP3 and I didn't notice any difference at all (probably because
POP3 is always slow).  Pretty slow is much better than not at all and
certainly is not useless.  I was able to strategically place a case fan
and then replace the shitty CPU fan on there with a new ball bearing one
with out even a reboot.  ...and performance returned to normal

Had the machine just turned off it would have been an immediate and
catastrophic loss of business communications (with pagers going off and
interrupting my other work).  But sure, if it was the machine in my
bedroom it wouldn't have mattered in the least and would have been a
minor annoyance.

Our point of sale terminals happen to run mostly AMD K6-2 450 CPU's on
an FIC board featuring the VIA mvp3 chipset.  What happens when the fan
totally fails on those?  Why the machine does nothing but bluescreen
over and over again.  Certainly not catastrophic in a business sense due
to the fact that there is more than one terminal available but still, a
pretty big annoyance to me.  Again, this is *ancient* hardware that
should have been replaced years ago.

I'm not trying to say that THG is "really awesome" or "all that and a
bowl of porridge" but to say that it's just blanket crap is going a bit
too far IMO.  Everything must be taken in context.  I'm also not just
another Intel fanboy.  But I'm not an AMD fanboy either.  I simply feel
that the AMD platform, especially now that Sun is involved, will only
continue to get better and better as enterprise features such as thermal
throttling are added, etc...  The next few years will undoubtedly be
very good for AMD as Intel has made many missteps recently.  This whole
prescott heat thing is on at least the same level as the i820 debacle.
You would think they would have learned from AMD when, in the past,
certain Athlon processors were burning up the bedrooms and required huge
(then) power supplies.  The P3 Tulutin owners sure were laughing then,
but now it swings the other way!

Sorry all for the long post.




More information about the lfs-chat mailing list