Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In my particular case(a laptop), a program (or the entire OS) might get stuck. And I have to reset.

Why would a processor fail anyway when it overheats? And what fails with these kind of symptoms?

The only thing I can come up with is that something melts. But that would sound irreversible, while if I turn off my overheated laptop, and turn it on again (after a few minutes) everything is still ok.

just wondering!

share|improve this question

migrated from Sep 22 '10 at 11:46

This question came from our site for system and network administrators.

If your processor exceeds it's safe operating temperature threshold, your computer should shut down automatically. I you want, disconnect the processor fan/heatsink from the actual processor die and start up the computer. It should only run for a few seconds before it overheats and restarts/shuts down. There should be protections in place to insure that your processor doesn't overheat and self-destruct. – Evan Plaice Feb 11 '11 at 1:19
up vote 8 down vote accepted

Typically every chip in a computer has a safe working temperature that, when exceeded can produce very strange results.

Every component in any type of integrated circuit will produce heat as a by-product of their operation and heat affects the performance of every last component. Resistors change their resistance very slightly when they get hotter, transistors leak more current as they get hotter and it is similar for every different type of component inside an integrated circuit such as a CPU. The problem is that in very large circuits all the tolerances of components can be very delicately balanced to produce specific circuit characteristics and when this balance is upset due to excess heat you can get changes in the way the circuit acts.

When the circuit characteristics change it can mean that wrong voltages get applied to crucial points and so the operation or the circuit can change. When this happens an operation that is invalid could happen and cause the processor to stall (as it is controlled by some very rigid logic) or output completely wrong data which the operating system or program is completely incapable of dealing with and so cause them to crash.

The reason everything is fine again after a reboot is that you have taken away the key problem, heat, and so the circuit is able to return to its correct operational state.

Prolonged excess heat can cause components to melt. Also, when circuits produce or sink excess currents or voltages due to functional changes caused by excess heat it can cause damage to them.

If you are regularly getting crashes due to excess heat then you need to either look at cleaning fans and so on or providing better cooling as the problem is only likely to get more frequent.

share|improve this answer
In addition to the correct info above, repeatedly heating/cooling anything will eventually cause cracking from thermal stress. – CarlF Sep 22 '10 at 16:18
thanks for a great answer – reinier Sep 27 '10 at 6:39

Determinism fails first. Basically, when the circuit paths get warmer, electrons are more likely to "jump over" to neighboring traces, producing unexpected results.

EDIT: Above is a popsci description of quantum tunneling. Additional effects should be increased resistance (because of the inreased heat) and increased heat (because of increased resistance).

EDIT2: this applies of course not only to the CPU, but generally to any component with very small feature sized ICs (RAM, Controllers etc.)

share|improve this answer
Also, this should be on, shouldn't it? – knitti Sep 22 '10 at 7:06
should it? Superuser has more of a software feel to it, than hardware questions – reinier Sep 22 '10 at 7:10
@knitti: electrons jump over because of heat?? is there a scientific name for this principle? – reinier Sep 22 '10 at 7:11
quantum tunneling – knitti Sep 22 '10 at 8:05
yes it probably should, superuser is for both software and hardware for individual users. serverfault is for sys admin problems. Especially when this question smells of homework rather than trying to solve an actual problem – JamesRyan Sep 22 '10 at 9:10

I doubt that it is actually your cpu overheating, most modern cpus have thermal protection long before you would see any effect. It is far more likely the chipset, graphics processor or memory, or power supply system.

For a practical solution there are several answers on superuser

share|improve this answer
@jamesryan: how would any of these overheating effects cause a complete stall of the OS? – reinier Sep 22 '10 at 9:29
If by complete stall you mean the screen freezes then it is likely to be the GPU. Often you will find other stuff is still going on and you arn't seeing it. The other main cause of locking up is disk access problems although these wouldn't be caused by overheating. Bad cpu or memory generally results in glitches or bluescreens. – JamesRyan Sep 22 '10 at 9:46

In general, electronics fail when heated because resistance to current increases when the conductor (or semiconductor) temperature is raised. When transistors on the microchip can't switch to "closed" because they're too hot, the CPU obviously can't function. (Transistors act as electrically-controlled switches.)

Also, the voltage from the power supply is basically fixed, so if the resistance of the components goes up, less and less current is able to pass through, and eventually the current drops to the point where it can't actuate microswitches and so forth.

Obviously all of the above applies to not just the CPU, but the GPU and any other support chips.

share|improve this answer
-1 Your info is a little off. With some components (like resistors) when they heat up the resistance increases. With others (like transistors) the resistance decreases with heat. When a component like a transistor decreases its resistance with heat their more current passes through the component increasing the heat more, decreasing resistance more until the unit burns out. The official term for this is called thermal runaway. – Evan Plaice Feb 11 '11 at 1:14

Sometimes, dust can also be the problem. I once had the same problem, but after a while, my POST failed.

I decided to open the laptop and see what was wrong. I live in a dusty area so, there was a lot of dust clogged in the heat-sink connecting both the processor and GPU:

enter image description here

Fortunately after cleaning every part of the laptop and applying some thermal paste under the CPU, it worked out.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .