Why Pentium?

[manny]

(e-mail address removed) writes:

I recall hearing somewhere that chip manufacturers stress-test chips
individually, then market the ones that won't run at high speed as
lower-speed submodels (rather like CCD manufacturers that test their
CCD sensors, selling the perfect ones to pros and putting the rest in
consumer gear).

It seems to me that most chip defects would be all or nothing--either
the chip runs or it doesn't--but perhaps there are things that can
affect individual chips without ruining them (?).

It's not all or nothing, and factory testing is used not only for
finding
defects but also for finding the maximum operating speed and
temperature ratings of each chip. And outright defects don't
necessarily mean the chip has to be dumped because they may
be sold as slower or lower capacity chips, an example being the
earlier AMD Durons, which were Athlons with defective L2 caches
that were partially disabled.

In any case, it makes me uneasy to stress my processor just to see if
it stops working.

Don't worry. Factory testing is done at higher temperatures than
you'll
ever use, unless your heatsink fan stops.
Don't worry

 

[manny]

The list is even shorter than that: all they care about is cost. Most
PCs today will do any type of business tasks that anyone might care to
perform, so clock speed, memory, and disk capacity are all
irrelevant--any PC will have more than enough of each for business
use.


You write off anything from Taiwan? I should think there are both
good and bad vendors. Don't some Intel motherboards come from Taiwan?

Taiwan brand electrolytic capacitors are aren't nearly as good as their
other products. This is why some Taiwan motherboards have Taiwan caps
for bypass but Japanese caps for their switching voltage converters,
and motherboards sold to Intel or Dell may have all Japanese caps.

 

[manny]

Mxsmanic wrote:

Well, for general office use but lets not exaggerate too much. There are
still plenty of things business does where computing power makes a
difference, like CAD, modeling, and animation just to name a few.

Just recently did a dual, dual core, system so the user would only have to
wait a few hours for their animations to render.

But that's not business but art, engineering, and science. The most
intensive business computing typically involves converting documents
to printable form.

I used to work for someone who gave the fastest, newest computers to
the executives and left his engineers doing CAD work on 100 MHz 486s.
Now I work for a much better person who gives the engineers everything
they want, sometimes even before they ask for it, and makes the
executives perform real work and get by with the slowest computers.

 

[Rod Speed]

That is a reason to more carefully scrutinize the failure
point, which was not an AMD processor but another
factor like fan or grease failure, chassis cooling problems.
One should never buy a system with the idea that one
of the basic fundamental needs will fail and thus Intel's
last-resort shutdown would matter.

Wrong. It makes a lot of sense to buy a system which wont
end up with a dead cpu if the cpu fan fails for whatever reason.
Just replace the cpu fan and carry on regardless if that happens.

Makes absolutely no sense to have to replace the cpu too
if something happens to the cpu fan or say the heatsink
clips give way or someone didnt install it properly.

Certainly that shutdown feature is better than NOT having
one, but it is not something that should be among primary
considerations in any remotely normal system, selection.

Its just another thing worth considering when deciding which
particular cpu to use, like that chipset question is too.

I suggest that you drew the wrong conclusion. A system
built with an Intel CPU but same problem the AMD one had, is
not trouble-free either. You saw the result of the problem
as a focal point instead of the cause.

Presumably he was pissed off about having to replace the cpu.

I would be too, particularly when that was one of
the most expensive components in the system and it
should have been designed better so that didnt happen.

Whatever that AMD CPU was, that it was a past generation
CPU is a sign that many alternatives from either manufacturer
produce more heat today, we can't just write-off the AMDs
as hot-running, and contrary to urban myth, many Intel
alternatives actually had a higher TDP but merely idled cooler.

Irrelevant to whether its bad design for the cpu to
end up dead due to something as trivial as a cpu
fan failure or bad installation of the heatsink etc.

Me too.

Put the $ towards the problem instead.

No thanks, I chose to buy the cpus that were designed better instead.

If it overheats the problem was the cooling system or maintenance
(lack of) towards cleaning out dust, replacing poor thermal
compound, or relubing junk fans (if for some reason it isn't
viable to replace them with good quality fans instead).

I still want a cpu that wont die if that stuff is neglected,
because its never possible to guarantee that that stuff
wont be neglected by someone who doesnt know any better.

Some have had trouble, for example Intel southbridge
USB issues/burnout. Pointing to one past chipset
used on AMD is no evidence against AMD itself.

No one ever said it was. I JUST said that one advantage
with an intel cpu is that you can have an intel chipset and
that on the whole there have been less problems with
those than there have been with VIA etc. Which is why
I choose to avoid VIA chipsets when thats feasible,
even when using an intel cpu.

Even in the past some chipsets for Intel posed
problems, like Sis 620 (or was it 630) refusing to
use UMDA for HDD on NT/2K/XP in many cases.

Which is another reason why I choose to use
Intel chipsets unless there is a good reason not to.

So long as the system proposed doesn't use the specific chipset,
there is no point considering that past generation chipset.

The problems seen with past generations of chipsets does
provide some evidence of the capability of that chipset supplier.

It brings up another prudent practice though, buying mature
platforms where there is ample feedback about issues.

Sure, I dont normally buy the bleeding edge, and choose to
avoid the chipsets which have demonstrated that the manufacturer's
design capabilitys have had severe problems in the past.

I do that with hard drives too.

That would be a silly random conclusion. One could
argue the same thing for an Intel Celeron w/Intel-
integrated video, as it is the most popular combination
for the highest selling market segment (OEM low-end).

Not at all. Many popular benchmarks make a ridiculous
assumption that one would only run a few of the premier
applications, newest versions of those. How many people
do you know that pay thousands of dollars every time a
newer version of their apps come out? Most people
don't, only getting newer versions when it happened to
ship with their new OEM system (which tends not to
have premier apps on it at all, except perhaps MS Office).

Take the typical apps of a few years ago and even Athlon XP
beat the P4 though online benchmarks suggested P4 beat it
most of the time (towards the end of the Athlon XP era at least).

There are reasons to choose either alternative, it would be
most valid to choose based on the specific, most common
or most demanding use the system will encounter...

In practice few personal desktop systems have
the performance determined by the cpu anyway.

as it is with a comparison of any two CPUs having different
architectures. Nothing wrong with a P4 or Pentium D where
it excells but the very specific use, not even a newer
version of the same application, must be considered.

Or realise that hardly any users would ever be able to pick
any difference any benchmark claims to see, with the main
exception being with games. And bugger all personal desktop
systems are used for demanding games anyway.

 

[paulmd]

As others have pointed out elsewhere in this thread, there's more to
life than sheer performance. As you say, they run cool, making passive
cooling a realistic option, and the case can be smaller too. I chose
a VIA chip when I wanted a small, quiet machine for my bedroom - I use
it for exactly the same tasks as my other machines and most of the time
I don't even notice the lack of power.

I had to reload the OS due to a virus infection on a C3 machine. It
was running at over 800mhz, but I swear a P2 400 would have beat it in
performance, no contest. Installing Adobe Acrobat 7 took FOREVER, I
wanted to get out and push. The customer told me she bought it new for
$400 sans OS. I think she overpaid. (I didn't tell her that )

I'm glad the VIA chip works for your situation, But I'd hate to inflict
it on an unsuspecting person. Lots of businesses have database systems,
or Lotus Notes, or various other intensive apps. If you're just running
Word, there's no problem.

 

[manny]

Nonsense. Executives are *very* careful with their pension plans.
ESRA and all that.

You're right, they're very careful with THEIR pensions, but the
pensions of their lower level employees are another matter.

Oh, that's why companies have CIOs; to read glossies.

The CIO of a company where I used to work didn't seem to do much
more when evaluating 401(k) plans because we ended up with a
plan that charged a minimum of 2.1% expenses. I phoned several
other plan management companies to see what they offered, all of
them were cheaper, both for the employees and employer, and one
even said they were too expensive for a company of our small size.

You read too much Tom's.

If the price was the same and I didn't need high CPU performance,
I'd have no objections to Intel P4-type chips.

 

[kony]

Hi there !

You mean you assembled the Intel Pentium WITHOUT thermal compound in
the first place ?????

Putting thermal compund on is one of the BASIC requirements when
assembling ... no wonder it ran HOT !!!!!!

Seems more likely the heatsink was installed with the OEM
thermal interface pad which "supposedly" eliminates the need
for thermal grease, even though many do find grease works
better.

 

[Bazzer Smith]

This was once the case. I always buy Intel, and I can tell you that the
last dual core processor at 3.2Ghz runs quite hot. In my case with
aftermarket (Arctic Cooler Freezer 7 pro) cooling, I have 50 degrees
Celsius (almost regardless of the ambient temperature).

Eh? A few here appear to be missing the point, what is I am saying that
AMD typicall consume less power.
You have 50 degree C ( I assume you mean chip temp, but you didn't
bother to say) because you expensive cooling system is doubling as a
space heater, hence you won't need much of a central heating system
as you PC is doing half of the job.

Your cooling system takes heat out of the chip(s) and pumps it into a
room (as does a fridge by the way).

 

[Bazzer Smith]

70C? No way. It probably takes at least 100C to actually damage the
silicon.

Err....if the chip is so hot that it is unstable then it is almost
certaintly
loweing the chips working life, OK it won't die overnight but it will
reduce its overall lifespan. So you find an appropiate temp to
run it at below a level which generates errors.
Yea you could run at 80 degrees C without damaging the
silicon in the short term, but you would get so many reboots
you might as well chuck it in a furnace for what good it would be.

 

[Bazzer Smith]

Hello,

I am in the market for a good computer, with a dual-core CPU. I keep
reading that Athlon is better than Pentium, Athlon is faster than Pentium,
and Athlon is lower priced than Pentium. But if that is the case, why do
most businesses have Pentium based PCs and not Athlon based PCs? Surely
most businesses research the pros and cons of a product before they make
their purchases. Thank you for clarifying this for me.

T.I.

You have seen the adverts have you not with the 'intel inside'
thing a then four musical notes dum dum-dum dum, this reprograms a humam
mind and subconsiously tells it to 'buy pent-i um'. (same 'sylables').
It's called brainwashing in marketting circles :O)

dum dum-dum dum

 

[kony]

As others have pointed out elsewhere in this thread, there's more to
life than sheer performance. As you say, they run cool, making passive
cooling a realistic option, and the case can be smaller too. I chose
a VIA chip when I wanted a small, quiet machine for my bedroom - I use
it for exactly the same tasks as my other machines and most of the time
I don't even notice the lack of power.

Some uses are fine with them. Car MP3 player, light duty
file or webserver are all good uses, and basic office, email
and websurfing (except elaborate scripts or animations) will
do fine too, but it is just so easy for a typical desktop PC
to be used for something more demanding too if it is a
primary use system instead of special-purpose.

Funny thing is, I've thinking about using a C3 platform
w/CompactFlash for a very quiet, low heat and power
fileserver. It will be a shame that I'm going to put hard
drives in it, more than doubling the power and making system
slightly audible, requiring a fan. ;-)

I've half a mind to see just how low power I can make it, by
underclocking, reducing voltage to the motherboard chips and
CPU I'm wondering if I can lower the whole thing to under
20W (sans the drives, RAID card and Gigibit nic). It's a
software raid card though, I'm suspecting I can't underclock
the C3 too far before it impacts drive throughput. Seems
like perviously I'd tested a similar Gigabit card and PCI
controller and found a Celeron 400 was about the lowest
performance level possible before the performance started
degrading, but then again I don't think it'll need any
particular level of floating point performance so the small
L2 cache may be the main issue as impacted by a lower,
underclocked memory bus speed.

I'm more eager to get my hands on an AMD Geode, but for the
time being the price hasn't dropped enough yet, the
aforementioned C3/motherboard combo I picked up for only
$35, and that it allows using a smaller PSU and less power
continuously running as a server, means it's practically
paid for itself already.

 

[Tony Hill]

Wrong. It makes a lot of sense to buy a system which wont
end up with a dead cpu if the cpu fan fails for whatever reason.
Just replace the cpu fan and carry on regardless if that happens.

If the CPU fan fails it is HIGHLY unlikely that either an Intel or AMD
CPU will need to be replaced. However, that being said, it CAN happen
with either one. I have definitely seen both AMD Athlon and Intel P4
chips fried from being run with a dead CPU fan. Usually this requires
the user to continue using the system for some time though after the
CPU fan stopped work, despite the fact that the computer would lock up
every few minutes.

Makes absolutely no sense to have to replace the cpu too
if something happens to the cpu fan or say the heatsink
clips give way or someone didnt install it properly.

If the heatsink clips give way you usually end up with one of two
situations, either your system won't boot at all because it will
detect that there is no CPU fan connected or your CPU will remain
stuck to the bottom of the heatsink when it gives way and the pins
will be damaged. Either way it doesn't make much difference if this
is an AMD or Intel CPU.

The real key is the second point you mentioned, what happens when
someone screws up while installing their heatsink. This is the one
situation where Intel's method of thermal protection tended to have an
edge over AMD's. The vast majority of cases I've encountered where
someone fried their AMD chip can be directly traced back to an
improperly installed heatsink.

Its just another thing worth considering when deciding which
particular cpu to use, like that chipset question is too.

A fairly minor one in my experience, unless you are not comfortable
with installing your own heatsink. Intel had a better heatsink
retention mechanism anyway which was more important in my experience.

Fortunately though, all of these issues are in the semi-distant past
now. With the Athlon64 AMD has a new and better heatsink retention
mechanism, they have the same sort of thermal throttling and shutdown
as a P4 and the nVidia and ATI chipsets for these processors seem very
up to par.

I still want a cpu that wont die if that stuff is neglected,
because its never possible to guarantee that that stuff
wont be neglected by someone who doesnt know any better.

Better not buy either AMD or Intel in that case, if you screw around
with the cooling mechanism enough either one will die on you.

Cooling systems are not something to just ignore when they fail hoping
that some short-term solution will save you. The real key idea behind
thermal throttling in Intel's P4 chips was that a system administrator
with an always-on server could get a warning about a failed fan and go
replaced within the next few minutes BEFORE the system crashed instead
of afterwards. This was definitely designed as a SHORT-term stop-gap
until the problem can be fixed, not a long-term solution.

In practice few personal desktop systems have
the performance determined by the cpu anyway.

As is usually the case, the CPU is just one of many pieces in the
puzzle. It's performance CAN have an impact though, even for personal
desktop systems. Mind you, the exact measure of "performance" is not
always something that a benchmark will show you. These days I'm
recommend dual-core chips to pretty much everyone looking for a
desktop system, not because they'll finish a benchmark faster, but
because they'll make the system much more responsive.

Or realise that hardly any users would ever be able to pick
any difference any benchmark claims to see, with the main
exception being with games. And bugger all personal desktop
systems are used for demanding games anyway.

The general rule of thumb is that most people will start to notice a
difference in performance once the benchmark shows at least a 20%
advantage. Before that one system may be faster, but most users will
have a tough time noticing it. Obviously this is not a particularly
scientific number or anything, just what many people have observed in
practice.

 

[kony]

Wrong. It makes a lot of sense to buy a system which wont
end up with a dead cpu if the cpu fan fails for whatever reason.
Just replace the cpu fan and carry on regardless if that happens.

Actually it is a sign that the person buying the system is
incompetent, and the one deploying it as well.

If even one moment is spent on considering the CPU's ability
to shut down while there were not good fans installed (which
make the risk of fan failure so remote as to be overshadowed
by any other reasonable risk), the effort was made in vain.
If there are good design decisions made towards cooling, the
CPU shutdown mechanism is of very little usefulness, far far
less than most other parameters in CPU selection.

Makes absolutely no sense to have to replace the cpu too
if something happens to the cpu fan or say the heatsink
clips give way or someone didnt install it properly.

"Something happens"?
If you can't keep "something" from happening, hire someone
who can. This isn't rocket science.

Same goes for heatsink clips and installation. Focusing on
the PROBLEM prevents downtime. Ignoring the problem is what
causes a perceived need for CPU shutdown features.

Its just another thing worth considering when deciding which
particular cpu to use, like that chipset question is too.

... about the last thing worth considering. Better than not
having it, but if you need it, the person who selected the
system and the builder/seller should be relieved of their
duties.

Presumably he was pissed off about having to replace the cpu.

He should have been more pissed off about why it happened.
As already written, if same thing happened in an Intel/P4
system we'd have to assume he'd be a little upset about that
too.

I would be too, particularly when that was one of
the most expensive components in the system and it
should have been designed better so that didnt happen.

Actually no. It is ridiculous thinking about the effect of
a problem rather than the source.

I could complain that a pad of paper burst into flames
because someone lit it on fire, but does it mean I should
buy flameproof paper or avoid tools that go around lighting
things on fire?

Irrelevant to whether its bad design for the cpu to
end up dead due to something as trivial as a cpu
fan failure or bad installation of the heatsink etc.

yes, it's "trivial" to do it right, so if the system weren't
in proper working order for long term use, the problem has
already occured, is not the future result.

Me too.


No thanks, I chose to buy the cpus that were designed better instead.

Then you choose to promote system downtime, failures. If
the system is designed properly the odds of the shutdown
feature being needed are too remote to be realistically
considered.

If you disagree, you have never bothered to learn proper
system component selection for long term use.

I still want a cpu that wont die if that stuff is neglected,
because its never possible to guarantee that that stuff
wont be neglected by someone who doesnt know any better.

Ok, if you presuppose a problem then that would help.
I'd rather presuppose the time should be spent on
eliminating the problem, or at the very least, checking for
this.

No one ever said it was. I JUST said that one advantage
with an intel cpu is that you can have an intel chipset and
that on the whole there have been less problems with
those than there have been with VIA etc. Which is why
I choose to avoid VIA chipsets when thats feasible,
even when using an intel cpu.

Ok, and again, it is pointless to name an entire company's
products rather than the specific one with the issue.
Someone could similarly claim "I had a p3 1.13GHz that
wasn't stable, this is proof we should never buy an Intel
CPU". It would be an equally invalid argument in the
context of system component selection today.

Which is another reason why I choose to use
Intel chipsets unless there is a good reason not to.

Ok, it's your $$. However, using them means you are
necessarily less informed through actual use of any
alternatives. IOW, you may then know a fair bit about them,
but not be able to reasonably contrast them to anything
else, _today_.

In practice few personal desktop systems have
the performance determined by the cpu anyway.

Possibly true, but we are talking about CPUs... which come
in different speed grades and corresponding prices. Quite
commonly people will spend more for a higher CPU # than
other system parameters so it is only reasonable to consider
what they get for the $.

Or realise that hardly any users would ever be able to pick
any difference any benchmark claims to see, with the main
exception being with games. And bugger all personal desktop
systems are used for demanding games anyway.

That's just it, the main difference is not just games. As
already written, you have to consider the app actually used,
not just the newest benchmarks of the newest apps. Likely
anything else, software evolves too, particularly for newer
CPUs the performance changes.

If one presumes a performance difference from a particular
CPU but without having the exact app and version they have
made an error, and likewise trying to draw conclusions about
similar tasks but still non-identical software. Games are
NOT the only place where AMD CPUs outperform "some" of
Intel's, it's merely one place where their raw performance
is shown, as it is in most apps not optimized for either
architecture. So pick your CPU then add onto it's cost the
cost for all the software you need to realize the benchmark
score.

 

[krw]

You're right, they're very careful with THEIR pensions, but the
pensions of their lower level employees are another matter.

Bullshit! They are responsible for the funds in the EMPLOYEES'
pension plan too. Look up ERISA. This is *very* serious business,
given the dollars in there.

The CIO of a company where I used to work didn't seem to do much
more when evaluating 401(k) plans because we ended up with a
plan that charged a minimum of 2.1% expenses.

You stupid twit! The CIO doesn't evaluate 401k's. Sheesh, grow up
already. (BTW, my employer's 401k charges 0.0%)

I phoned several
other plan management companies to see what they offered, all of
them were cheaper, both for the employees and employer, and one
even said they were too expensive for a company of our small size.

Which is irrelevant to the nonsense you're spouting.

If the price was the same and I didn't need high CPU performance,
I'd have no objections to Intel P4-type chips.

You're reading too much Tom's. (...or perhaps RWT).

 

[paulmd]

Some uses are fine with them. Car MP3 player, light duty
file or webserver are all good uses, and basic office, email
and websurfing (except elaborate scripts or animations) will
do fine too, but it is just so easy for a typical desktop PC
to be used for something more demanding too if it is a
primary use system instead of special-purpose.

Funny thing is, I've thinking about using a C3 platform
w/CompactFlash for a very quiet, low heat and power
fileserver. It will be a shame that I'm going to put hard
drives in it, more than doubling the power and making system
slightly audible, requiring a fan. ;-)

You could use laptop drives.

 

[kony]

You could use laptop drives.

Sure, but that kills the capacity per $ and the performance.
I don't really care so much about the noise as it isn't
going to be sitting within earshot but having the fans means
that someday I'll have to pop off the filter panel and clean
or replace it. It's not much of a concern though, I already
have a couple other filesevers with large filters over the
entire 5-1/2" bay area and from the slow dust accumulation
it looks like they could go several years inbetween filter
replacement intervals.

 

[kony]

Irrelevant to whether its bad design for the cpu to
end up dead due to something as trivial as a cpu
fan failure or bad installation of the heatsink etc.

You are overlooking that a gradual overheating situation
with either AMD or Intel CPUs, has an overheat shutdown
mechanism in place. Since Intels' was CPU-integral earlier,
those CPUs have a marginally better protection but in
practice you would have to have a rather unlikely overheat
scenario- not as slow as if the fan failed, but not as fast
as if the heatsink came off.

If the heatsink installation was bad such that it didn't
make contact, the clamp came off or whatever, the system can
still fry a P4. It has been done, a P4's shutdown mechanism
cannot respond fast enough to counter the rapid rise in temp
from cold-off to on-without-heatsink-contact.

Citing one example of an old platform with an ineffective
means to power off a system with an Athlon in it is a
similar situation to any other past era issues- unless you
are buying that particular old tech, it is non-applicable to
parts selections today.

 

[Merrill P. L. Worthington]

I own two PC's I put together myself, one with the Athlon 2600+, the other
with the Intel Pentium D dual core. The latter actually runs a bit
faster(even though both PC's have the WD SATA Raptor drives running at
10,000 RPM). But it also runs at a higher temp and after nine months was
beginning to overheat a bit. So I applied some thermal compound the other
day which has corrected the problem.

I'm surprised that the Intel only runs "a bit faster" than the old
Athlon. That doesn't speak very well of the Pentium D.

 

[Rod Speed]

If the CPU fan fails it is HIGHLY unlikely that either an Intel or
AMD CPU will need to be replaced. However, that being said,
it CAN happen with either one. I have definitely seen both AMD
Athlon and Intel P4 chips fried from being run with a dead CPU
fan. Usually this requires the user to continue using the system
for some time though after the CPU fan stopped work, despite
the fact that the computer would lock up every few minutes.

And that isnt that uncommon with the average user.

If the heatsink clips give way you usually end up with
one of two situations, either your system won't boot at all
because it will detect that there is no CPU fan connected

Not if it fails in a way that sees the heatsink just
loose so it doesnt cool the cpu properly anymore.

or your CPU will remain stuck to the bottom of the heatsink
when it gives way and the pins will be damaged.

That assumes it actually ends up reefing the cpu
out of the socket. That doesnt normally happen.

Either way it doesn't make much difference if this is an AMD or Intel CPU.
Wrong.

The real key is the second point you mentioned, what
happens when someone screws up while installing their
heatsink. This is the one situation where Intel's method
of thermal protection tended to have an edge over AMD's.

It isnt the only one.

The vast majority of cases I've encountered where
someone fried their AMD chip can be directly traced
back to an improperly installed heatsink.

Yep, which is the reason that I avoided amd
cpus while they had no thermal protection.

A fairly minor one in my experience, unless you are
not comfortable with installing your own heatsink.

It isnt the only one.

Intel had a better heatsink retention mechanism
anyway which was more important in my experience.

Yeah, and avoided that bodgy approach the early
athlons had with a small easily damaged cpu top etc.

Fortunately though, all of these issues are in the semi-distant
past now. With the Athlon64 AMD has a new and better
heatsink retention mechanism, they have the same sort
of thermal throttling and shutdown as a P4

That was already covered earlier in the thread.

and the nVidia and ATI chipsets for
these processors seem very up to par.

Time will tell, particularly with the latest amd cpus.

Better not buy either AMD or Intel in that case, if you screw around
with the cooling mechanism enough either one will die on you.

Its a lot harder to **** that up with an intel cpu up till now.

Cooling systems are not something to just ignore when
they fail hoping that some short-term solution will save you.

Yes, but its also very undesirable to have the most expensive
component in your system get killed by something going wrong
when it should have been designed to handle that gracefully.

Which is why even amd does that now.

The real key idea behind thermal throttling in Intel's P4 chips was
that a system administrator with an always-on server could get a
warning about a failed fan and go replaced within the next few
minutes BEFORE the system crashed instead of afterwards.

Doesnt explain why celerons had it too.

This was definitely designed as a SHORT-term stop-gap
until the problem can be fixed, not a long-term solution.

Wrong, most obviously with the mobile intel cpus.

As is usually the case, the CPU is just one of many pieces in the
puzzle. It's performance CAN have an impact though, even for personal
desktop systems. Mind you, the exact measure of "performance" is not
always something that a benchmark will show you. These days I'm
recommend dual-core chips to pretty much everyone looking for a
desktop system, not because they'll finish a benchmark faster,
but because they'll make the system much more responsive.

Thats arguable given what most do on their personal desktop systems.

It can make a lot more sense to spend that money on ram instead.

The general rule of thumb is that most people will
start to notice a difference in performance once
the benchmark shows at least a 20% advantage.

Thats overstating it with systems that arent used for gaming
or say crunching video files from one format to another.

They're about the only things where the benchmark performance
is even visible with most personal desktop systems now.

Before that one system may be faster, but
most users will have a tough time noticing it.

And after that with most personal desktop systems too.

Obviously this is not a particularly scientific number or
anything, just what many people have observed in practice.

But not with the way most personal desktop systems are actually used.

 

[David Maynard]

I'd expect silicon chips to last long enough that their MTBF is to all
intents and purposes irrelevant - most solid state electronics has a
pretty good lifespan - the only thing that tend to let the side down
are electrolytics.

Well, silicon chips do seem to be stunningly reliable but they can and do
fail or else there'd be little use for the entire semiconductor failure
analysis business.

Whilst obviously we can't discuss modern machines in this context, as
a case in point I have an old Apricot LS Pro (something of a
collectors item now) powered by a a Cyrix 486SLC. This is a tiny,
_plastic_ chip with no heatsink and no fan. It was in daily use until
I picked it up maybe 18 months ago. I've no idea how hot the chip
gets - far too early for a motherboard sensor - but it's too hot to
touch in operation so that's probably at least 60C. It still works a
charm, maybe 14 years after it was built. Most people won't be using
the machine they have today in that amount of time, unless of course
like me they have a interest in old kit.

Yes, and I've got numerous oldies but goodies as well. How that translates
into modern processor with much higher current densities is another
question, and one I don't have the answer to since the data isn't public.

The issue of 'what temp?' should be, to some extent, almost automatic as
the system should be designed to operate properly over the entire
'operating temperature range' with 'normal' operation usually significantly
less.

I don't think that necessarily holds, however, for systems employing
aggressive thermal management/throttling techniques because the processor
can be operated closer to the maximum limit under 'normal' ambient
conditions and throttled back in the higher temperature environment.