Why Pentium?

Nope, its just a basic recognition of how many users operate.

Thats a completely silly claim.

Wrong with most personal desktop systems which dont need any
careful choice of cpu at all. What happens in reality with the absolute
vast bulk of those is a choice based on value for money instead.

It makes a hell of a lot more sense to have the cpu
behave gracefully when the shit hits the fan instead.

Have fun explaining how come even amd
now have a decent shutdown mechanism.

Its nothing like as black and white as you are claiming.

Nope, not if its a value for money personal desktop
system where there are no considerations in the
choice of the cpu except whats currently decent value
for money with minimal hassles with chipset quirks etc.

You clearly dont have a clue about what value
for money personal desktop systems are about.

It wouldnt have happened if it was just an awkwardly
designed heatsink which could be installed improperly.
He would have had the chance to see that the cpu temp
was way out of line and check the heatsink to see why.

Its completely mad that an awkwardly designed heatsink
should see the cpu dead, particularly with the amd cpus
that didnt at that stage come with boxed HS and fan.

Mindlessly silly.

Mindlessly silly.

Nothing like the same thing as an awkwardly
designed HS which can be improperly installed.

Gets sillier by the minute.

Gets sillier by the minute.

Gets sillier by the minute.

Not a clue, as always.

There are ALWAYS problems.

Have fun explaining why even amd now has thermal shutdown.

No its not, its evidence of how that company does things.

Gets sillier by the minute.

Doesnt cost me a cent.

What I choose to buy for myself is an entirely separate
matter to what I have seen problems with.

Gets sillier by the minute.

Or it makes much more sense to buy on the basis of value
for money instead with systems known to be reliable.

Never said it was.

In practice with what is done on most personal desktop systems,
benchmarks are completely irrelevant. The user wont even notice
any difference between any of the sensible alternatives in practice.

And most users of personal desktop systems wont even notice if they
arent running demanding games or stuff like transcoding video files.

In practice with what is done on most personal desktop systems,
benchmarks are completely irrelevant. The user wont even notice
any difference between any of the sensible alternatives in practice.

In practice games and transcoding video files are about
the only things that are done much on most personal desktop
systems where you will notice any effect of the cpu at all.

Or just pick what's best value for money instead.

That sort of benchmarking is a complete wank in the real world.

 

No I'm not, thats just one area where there
isnt anything in it between intels and amds.

Nothing marginal about it when the cpu fan
fails or the heatsink isnt installed properly.

Wrong again. Its quite common with fur buildup on the cpu fan.

And if the thermal pad has just got damaged, it may well handle it fine.

Duh, obviously since the current amds do now have a decent thermal shutdown mechanism.

He was clearly commenting on the downsides of not having a decent
thermal shutdown mechanism, and amd clearly NOW agrees with him.

 

Cost is one major factor, considering notebook drives tend
to be around $1 per GB still and the proposed fileserver
will have at least a TB mirrored. Then there's the 40/44
pin adapters dangling in air behind the drives, something
I'd like to avoid for reliability's sake.

I'd expect the notebook drives to be the most significant
bottleneck, that I can underclock/undervolt the CPU and
mainboard and still have over 30MB/s with desktop drives.
It won't be necessary to underclock it but can't hurt in
achieving long lifespan either, the system can remain viable
for at least 10 years on GbE and a PCI based PATA & SATA
card(s). Whatever I come up with for AC power conditioning
will probably cost more than the entire system (sans drive
costs). Plus, if I keep power consumption below a certain
threshold, I have plenty of parts for building a redundant
power board. Beyond that threshold, it becomes yet another
expense, and same situation with some UPS I already have.
I could justify the expenses if necessary, if I didn't have
any fileservers yet but I do, so... It becomes more a matter
of a creative exercise during a routine pre-planned
obsolescence of one of my older fileservers. They run fine
but I'd rather replace while they work rather than after one
had failed, especially when it can be done at my leasure
instead of an immediate necessity. That just allows playing
around a bit before it's done.

Essentially I'll be leveraging the most cost effective
modern technologies and ignoring those with a high
cost:benefit ratio. I can underclock to lengthen service
intervals (except the inevitable drive failures over several
years use) and promote longest motherboard lifespan. A
light duty filesever just doesn't need much in the way of
CPU performance and the software raid may be an issue but on
a similar existing box with a Celeron 500, it never reaches
100% CPU utilization more than momentarily, fractions of a
second with averages (during transfers) well below 75% (I
forget the exact percentage at the moment).

If higher CPU performance had significant gain I could just
put a higher performance platform to use but I expect the
biggest bottleneck to be the 32bit/33MHz PCI bus as it was
the case with post-1GHz Tualatin fileservers and it would
raise costs by an order of magnitude to overcome that PCI
limitation with a newer PCI Express or ported GbE platform &
complimentary CPU. I'll have to benchmark it, maybe it
won't perform well enough underclocked but it's hard to
speculate as too few people have benchmarked C3 in these
kinds of uses. I'd seen some filesystem benchmarks that
make C3 look fine but not with software raid. So I'll
underclock then compare and see...

 

Isn't anything in what?

Either platform has had overheat shutdown for years now.
Early socket A didn't but were you buying an early socket A
based system new today?

Marginally better means that with the thermal sensor in the
CPU, it will react faster, but it need not react that fast
if the fan failed or 'sink was clogged with dust, because
the temp doesn't rise so fast in these conditions.

Yes and that scenario you post IS handled by AMD's solution.
Did you think there was NO thermal shutdown at all? Perhaps
this is where you are mislead.

So how do you propose to damage it?

Again this is a very narrow change in thermal rise, "IF" it
were too fast for one thermal sensor to handle shutdown but
still slow enough for another to do so safely.

.... and the XPs did too on the motherboard. Fan fails?
Motherboard shuts it down. Dust? Again, motherboard shuts
it down. Heatsink falls off? CPU may fry but so have P4s.

You have to reach to find a realistic scenario where it'll
make a real-world difference.

He was commenting about it as if it's a reason not to choose
AMD, now. Instead, it is a reason not to choose that one
old platform, which is same thing I've commented on all
along, that a particular issue with some old platform is not
an indictment against an entire company's line of products,
particularly later generation products. Again, same thing
applies to any earlier generation Intel bugs, it in now way
reflects on what you'd buy today as a current gen. part.

 

On Wed, 5 Jul 2006 14:32:15 +1000, "Rod Speed"

Is this why you need backup systems in place?

For over a decade systems managed to run without CPU
shutdown mechanisms, and do so fine so long as the fans
worked and dust was kept in check. It's really that simple,
elimination of the failure points is what makes a good
system, not trying to fail gracefully then having system
downtime till the failure point is fixed.

Again, I didn't claim a CPU shutdown mechanism is worthless,
rather than it should never be considered until after "there
were ... good fans installed". One is necessary for a
long-term reliable system, the other only if the system
wasn't reliable in the first place.

 

With cpus that dont burn up if the cpu fan fails.

And few bother to do a damned thing about dust.

More mindlessly silly stuff. Anyone with a clue designs a
system so that the most expensive component in the system
doesnt die if something as basic as a cpu fan fails.

More fool you. Plenty, including intel and now amd, have enough of a
clue to ensure that the the most expensive component in most systems,
the cpu, doesnt die when something as basic as the cpu fan dies.

More mindlessly silly stuff.

Just as well that you dont get any say on the design of anything, ever.

 

You cant actually be THAT stupid.

Anything in the brand of the cpu, stupid.

So much for your stupid pig ignorant claim that that has no value.

Irrelevant to what was being discussed, whether cpus that
will die if the cpu fan stops, should have overheat shutdown.

Have fun explaining why even amd has that now.

Or perhaps not.

Having fun thrashing that straw man are you ?

By taking the heatsink off and replacing it again.

Have fun explaining why even amd has that now.

Have fun explaining why even amd has thermal shutdown now.

Have fun explaining why even amd has thermal shutdown now.

No he wasnt.

Have fun explaining why even amd has thermal shutdown now.

 

Yes, it's much more preferable to have your computer shut down without
notice, and lose any work in progress.

(*rolls eyes*)

 

You'd only lose more than 10 minutes or so work.

Decent software autosaves periodically.

Backup software configured correctly monitors changes in document
folders, would see the autosave and back it up.

So what you're actually indicating by that comment is that you're quite
incompetent.

--
Conor
Sig under construction. Please check back when Duke Nukem Forever ships
and/or Windows Vista is released.

Cashback on online purchases:
http://www.TopCashBack.co.uk/Conor/ref/index.htm

 

Huh? I would think that an engineer/artist in the business of
producing CAD/animation would consider "that" a business activity. A
lot of people use the same systems for more than one role especially
for small businesses. Furthermore, large spreadsheets can take a
pretty long time to update with every change (I've seen some that
takes like over a minute for each refresh).

 

WORK IN PROGRESS.....get it? 10 15 minuters... whatever its WORK IN
PROGRESS..... and hang around a video rendering machine and listen to the
lab SCREAM when a system goes down......

 

I claim there is no value in considering it until after the
problems you cite are addressed, the REASON it would be even
slightly useful.

Again, that is not what I was discussing, rather the
incompetence of someone who gives this a thought before
setting up a system properly, reliably.

.... because some people who shouldn't be building systems,
do. Same people tend to make other mistakes as well, and in
the end AMD and Intel took a step that guards them against
some forms of incompetence but the system itself still
suffered the downtime.

 

Maybe if you lose enough you'll learn to just set a system
up right instead of trying to shift the burden?

 

You routinely run large jobs on systems not proven to be
stable? Find a mirror when it's time to scream.

 

On Wed, 5 Jul 2006 20:14:38 +1000, "Rod Speed"

You are actually suggesting that someone is SO reckless that
they're buying an expensive CPU for a box they arent'
bothering to fit with good fans?

Apparently so, as I did not argue there was no benefit at
all to the shutdown mechanism, rather it is an unrealistic
concern in a properly configured (from a hardware
standpoint) system. I argued it shouldn't be considered
BEFORE you had the system set up properly against the kinds
of problems you suggest would make the feature of benefit.

Note the word "before" above. You argued against that, and
in doing so, show no clue about system uptime. A down
system in itself can cost more than the loss of the CPU,
even multiple times as much.

 

That's a very limited view of "business computing". As a broad category,
e.g., decision support systems can do some very heavy duty calculations,
whether it be financial analysis or strategic & tactical planning for any
part of a manufacturing business.

 

Have fun explaining why even amd now has thermal shutdown.

Have fun explaining why even amd now has thermal shutdown.

Have fun explaining why even amd now has thermal shutdown.

Lying now.

Never ever could bullshit its way out of its predicament.

Have fun explaining why even amd now has thermal shutdown.

 

Have fun explaining why even amd now has thermal shutdown.

You have always been, and always will be, completely and utterly irrelevant.

Have fun explaining why even amd now has thermal shutdown.

Wrong, as always.

Or they actually have enough of a clue to realise that it makes
absolutely no sense for the cpu to die when something as basic
as a fan failure can do that to the most expensive component,
particularly when thermal shutdown costs peanuts.

Just as well you get no say what so ever on anything at all, ever.

 

Have fun explaining why even amd now has thermal shutdown.
Because cooling fans fail. (duh) -Dave