Question about CPUs. (Warning: A stupid one)

[coldfish]

I remember reading a piece about "proton drift" or something like that, talking about how the electron flow can actually move protons over time.  Its becoming more of a problem as dies shrink further i believe.

Now, where's that link?

[bloodline]

coldfish wrote:
I remember reading a piece about "proton drift" or something like that, talking about how the electron flow can actually move protons over time.  Its becoming more of a problem as dies shrink further i believe.

Now, where's that link?

I think you mean Electromigration, which happens when you over current a circuit... older CMOS CPU's seem to suffer from this effect.

link: http://en.wikipedia.org/wiki/Electromigration

[Xamiche]

Only few chips are actually capable of downclocking when overheated (e.g. Pentium 4). Most simply crash.

It not a matter of downclocking. You still lose performance when it over heats because of errors in calculations, or rather operations, and time spend doing the operation again because of the error. Of course if the chip continues to over heat, it will eventually crash. Don't just take my word for it, try it. Take a working Pentium system, run a first person shooter with a frames per second read-out and then slow the CPU fan down. Watch the frames per second drop. We did this in our microprocessor fundamentals unit. I say use a Pentium chip because it more resilient to heat than an AMD chip. It will survive where an AMD will most likely fry.

[Sparky]

JJ wrote:
Stop........De-fraging a hard-drive is pointless.  It will take hours and hours, and you will not notice the real world difference (apart from the old its supposed to be faster so it is effect).  you will never gain back the time lost in de-fraging in a performance gain.  And guess what as soon as you have de-fragged the drive it gets fragmented again.  

Ummm ... defragging is a "good thing"!

If your drive is getting fragmented really quickly then you are probably messing with lots of little files or running an app that deals with lots of little bits of files (Bittorrent and P2P apps are good ones for fragmenting a drive quickly) .. to ease the pain a bit you could get the apps in question to work on another drive (rather than your main boot drive), or you could look at reformatting your drive with a different block size, gets to be a bit of a balancing act when you do that though (wasted space vs performance etc).

Just my NZ$0.02 :-)

[mikrucio]

well I work in a Computer repair store.. and I can tell you from 7 years of experience. Hard disk drives can
and WILL slow down from 30mb/s to about 1mb/s for example.
NOT ALL hard disks will do this. some models are more prone than others... so there is truth to the above posts...
Just like ANY electrical goods. components WILL fail over time.. And as far as CPU slow down over time is concerned..
Yes it IS possible for reasons mentioned above... socket A Athlons were VERY prone to this..... Again it wont happen to ALL CPU's which is what this post is about.. In fact the CPU in a computer will outlast every other component.. and are ALOT less bound to fail.

[asian1]

Self repairing Power and PowerPC CPU from IBM:

http://www.theregister.co.uk/2004/08/02/ibm_efuse/

[Zac67]

It not a matter of downclocking. You still lose performance when it over heats because of errors in calculations, or rather operations, and time spend doing the operation again because of the error.

This is complete nonsense - how's the CPU supposed to know when it's wrong? There's no way to do it other than do everything twice - and nobody's designing that way (not in consumer world).

Take a working Pentium system, run a first person shooter with a frames per second read-out and then slow the CPU fan down. Watch the frames per second drop.

Yes, a P4 will downclock when it's sensor shows overtemp - not because of 'errors in calculation'. A lot of calculation done by CPUs is that of addresses and jumps, when something goes wrong there, there's nothing to do for recovering. Either the program trashes data or crashes.

I say use a Pentium chip because it more resilient to heat than an AMD chip. It will survive where an AMD will most likely fry.

Even more nonsense. You should have paid more attention or maybe your teacher has told you wrong: take a PIII and stop the fan. Voilá: crashed. Again: P4s monitor their temp and skip clock cycles (on purpose) when overheated.
AMD were said to fry easily because the Socket A generation had the heatsink directly attached to the naked die. When the HS was not fixed accurately, the chip fried. This does not happen any more with Socket754/939/940/AM2 CPUs.

[Piru]

@mikrucio

well I work in a Computer repair store.. and I can tell you from 7 years of experience. Hard disk drives can and WILL slow down from 30mb/s to about 1mb/s for example.

They don't actually slow down per se. The slowdown is from read/write retries, or from the extra seeks resulting from bad sector/track remapping.

To check if the HDD has any such blocks, check S.M.A.R.T attribute 5 (Reallocated Sectors Count). Also attributes 196, 197 and 198 should be checked for.

Regardless, the actual physical disk access is exactly as fast as before. The areas that are not affected by the bit rot work at full original speed, for example.

That being said, it needs to be really severe case before it begins to affect the performance. Couple of dozen of remapping sectors will not affect performance, really.

NOT ALL hard disks will do this. some models are more prone than others.

Some HDDs are much more prone to failure, yes.

[Xamiche]

Enough, I don't want anymore forum wars. I can go to the Archlord forums for that. I know what we observed. When we slow the fans the programs performance degraded.

I never said a CPU can get slower with age. I never said P4 don't down clock. My post is a response to someone who wondered why they benchmarked lower with the same system several years apart. Our trial may cast some light on why.

[coldfish]

Right on Bloodline, thanks for the link.

Proton drift, bah, I must've been sleepy when I wrote that.
  :-D

[Fats]

Xamiche wrote:
Enough, I don't want anymore forum wars. I can go to the Archlord forums for that. I know what we observed. When we slow the fans the programs performance degraded.

But you made the wrong conclusion from your observation. The CPU starts to run slower in order to generate less heat because it was especially designed to do so when a temp sensor senses it is becoming too hot.
It doesn't run slower by itself because a hot CPU chip runs slower.

greets,
Staf.

[PulsatingQuasar]

Depends how you look at this but a graphics processor like the Geforce 6800 has pipelines that are part of it's speed. So if a couple of pipelines should breakdown over the course of a few years you could turn off those pipelines and still have a working card but slower.

This is basically the difference between several models of the 6800. They are just missing some pipelines.

[AmigaMance]

Fats wrote:

Xamiche wrote:
Enough, I don't want anymore forum wars. I can go to the Archlord forums for that. I know what we observed. When we slow the fans the programs performance degraded.

But you made the wrong conclusion from your observation. The CPU starts to run slower in order to generate less heat because it was especially designed to do so when a temp sensor senses it is becoming too hot.
It doesn't run slower by itself because a hot CPU chip runs slower.

 I have read every post on this thread and now that i understand how things works i'll have to agree with you. There is another mystirious issue related to this though:
 I've read 3 posts from 3 different users, here and in other forums who claim that their 68040 becomes slower when it is not properly cooled. Not 1 or 2 but 3 reports.
 Note: I haven't experience anything similar as an 68040 owner, but my CPU is always cooled actively and it runs at 25mhz only.
 I know that 68040s are the most hot chips of the 68k family etc, etc, but still, this doesn't explain what these users experienced. It should only make their CPUs less stable, not slower.
 You could classified it as placebo but they seemed quite confident about it from their posts.

[Zac67]

Depends how you look at this but a graphics processor like the Geforce 6800 has pipelines that are part of it's speed. So if a couple of pipelines should breakdown over the course of a few years you could turn off those pipelines and still have a working card but slower.

You don't have the option to deactivate single pipelines without altering the driver (depending on design, modding the driver may not be enough). If one of the pipelines fails, you'd get garbled output and bin the card.

This is basically the difference between several models of the 6800. They are just missing some pipelines.

True. But the different version are detected by the driver and programmed accordingly. If there are no hidden tweaks in the driver and no way to simulate another version by opening/closing jumpers, reflashing BIOS etc., there's no way to make the card stop using the broken pipes.

[da9000]

Xamiche wrote:
Enough, I don't want anymore forum wars. I can go to the Archlord forums for that. I know what we observed. When we slow the fans the programs performance degraded.

I never said a CPU can get slower with age. I never said P4 don't down clock. My post is a response to someone who wondered why they benchmarked lower with the same system several years apart. Our trial may cast some light on why.

You also compared oranges to apples. A Celeron of that age (466Mhz) has no throttling like a P4 in your tests. Unrelated. Just listen to Piru and Zac and those who agree with them. They know what they're talking about.

[da9000]

AmigaMance wrote:
 I've read 3 posts from 3 different users, here and in other forums who claim that their 68040 becomes slower when it is not properly cooled. Not 1 or 2 but 3 reports.
... You could classified it as placebo but they seemed quite confident about it from their posts.

That maybe so, but who's to say that those users have:

1) made the right conclusion (I can see a speed increase when I muck around with hardware, but along with adding a fan, I also swapped video cards, and perhaps due to some intricate timing issues which I didn't consider due to lack of knowledge [think of Shared-Memory video solutions on PeeCees] I saw a speed increase, therefore my conclusion that adding a fan makes things faster is flawed)

2) are qualified to make the right conclusion (many people can change a fan, but few understand the more or less complete internal workings of a computer to such a degree as to make qualified conclusions. In fact, what's evident from this thread is who is a low level programmer / systems engineer and who is not. I can quite clearly give you a run down of the names without ever knowing anything about these people, and it's simple: a low level programmer / system engineer would swear by what Zac is saying as far as instructions and data in CPU calculations: there's no "error correction" and "retries" and such B.S., other than CRC/ECC calculations on the data buses of course. The CPU either works or doesn't, which will produce a hardware exception, or a Guru Meditation if you will. To think of a modern day CPU as having error corrected address/offset/pointer/arithmetic calculation is simply ludicrous)

Synopsis:

CPUs don't slow down as they age. They die a sudden death.

Hard disk drives and other mechanical devices don't slow down as they age, per se, but can exhibit slower "observed"  or "net" behaviour, which is due to bad media (bad blocks) or damaged mechanical parts (motor spins slower, possibly due to power issues, or due to friction caused by heat or cold, physical shock, and other such mechanical failures)