When OVERCLOCKING, how does an increased VOLTAGE harm the CPU?

[x86]

<<
The indirect source of problems from voltage is, as many others pointed out, heat. Heat affects all of the above issues strongly. They would all have an increased effect if somehow the die temperature were a constant and the voltage were increased, but heat is an additional factor. So increased heat leads to these effects occuring more quickly.

Patrick Mahoney
Microprocessor Design Engineer
Intel Corporation >>



PM - If the normal opperating temperature of the CPU were to be maintained and the voltage were to be increased, will the degrading effects still occur?

Thanks

-DocSmarts

 

[pm]

<< In addition, PM stands for Private Message or Private Mail, so pm would probably prefer to be referred to by his lowercase handle of pm. >>

There wasn't "Private Messaging" when I joined, so about the only downside of my nick at that time was that it made it hard to search for my posts (since half or more of the posts contain "pm"), but nowadays, it is a little confusing with "PM" and "pm".


<< PM, does overclocking without changing the voltage affect the lifespan of a CPU or GPU? Does the faster switching cause increased wear on the CPU? >>

Yes, but nowhere near the effect on increasing the voltage. Voltage increases result in an order of magnitude or more decrease in reliability/operating lifetime. Increasing the frequency increases the heat of the CPU (and heat affects all of the previously mentioned reliability issues), and it increases the current on the power wiring leading to increased electromigration issues. But like I said merely increasing the frequency is a far less significant problem from a reliability standpoint than increasing the voltage.

Since someone will probably mention it sooner or later, the obvious question is "by how much does increasing the voltage by xx% reduce the operating life of a CPU?". This is a difficult question for me to answer both because I honestly don't know, and if I did I couldn't say. There can be no doubt, however, that increasing the voltage of the CPU leads to an increased probability that it will burn out on you sooner than it would otherwise. I seem to recall seeing a lot of posts of dead Celeron 366's on the forums around that time as well.PM - If the normal opperating temperature of the CPU were to be maintained and the voltage were to be increased, will the degrading effects still occur?


<< PM - If the normal opperating temperature of the CPU were to be maintained and the voltage were to be increased, will the degrading effects still occur? >>

This would be a difficult trick to do since the power dissipation of the circuitry increases by a square of the voltage increase, the simple form of the power dissipation equation is:
Power = Frequency x (constant which is the capacitance of the CPU circuitry) x Voltage ^ 2

Since power increases by the square of the voltage, it is hard to increase the voltage and maintain a constant temperature. But even if you could, as I mentioned above, increasing the voltage increases the effect of all of the reliability mechanisms that I mentioned.

 

[Barnaby W. Füi]

hhahaha pm puts the proverbial smack down on the kiddie moderators and they change the doc's name back.

ROFLMAO!!

 

[x86]

<< Yes, but nowhere near the effect on increasing the voltage. Voltage increases result in an order of magnitude or more decrease in reliability/operating lifetime. Increasing the frequency increases the heat of the CPU (and heat affects all of the previously mentioned reliability issues), and it increases the current on the power wiring leading to increased electromigration issues. But like I said merely increasing the frequency is a far less significant problem from a reliability standpoint than increasing the voltage. >>



If the electromigration usually increases as the clockspeed increases, what is preventing higher clocked CPUs from being harmed more than lower clocked CPUs manufactured under the same process?

Thanks

-DocSmarts

 

[x86]

<< hhahaha pm puts the proverbial smack down on the kiddie moderators and they change the doc's name back.

ROFLMAO!! >>



I think not!

You are acting quite childish, please stop.