Is heat the killer or voltage itself?

[Devlafford]

In overclocking, does the lifespan-reducing effect come from simply upping the voltage itself, or is it from the temperatures created by increasing power consumption?

In addition, can someone explain power phases to me? My board has 8+2, I've got an A10-7850K running fine at 4.1 without touching voltages or Load Line Calibration or anything else. I don't understand what the power phases have to do with anything.

It's being cooled by a Cooler Master Seidon 120V, thermal margin stays above 50C currently.

[Lays]

Both.  The idea would be the more voltage you push, the safer it will be to keep it at a lower temperature.

There's an article somewhere I'll see if I can find it, it explains completely how chip degradation works, it's super interesting actually.

The jist of the idea would be, 1.4V 24/7 at 50C would be safer vs 1.35v @ 70c.

If anyone can help me find the article, I think I remember it being on Tom's hardware.  It may of been a post somewhere, I wish I would of bookmarked it. -.-

[WinNut]

Each power phase provides more power, this is the case so that resistors aren't needed when the CPU wants to downclock at idle.

[Marvzl1357]

Both.  The idea would be the more voltage you push, the safer it will be to keep it at a lower temperature.

There's an article somewhere I'll see if I can find it, it explains completely how chip degradation works, it's super interesting actually.

Is it this one? http://spectrum.ieee.org/semiconductors/processors/transistor-aging

[Lays]

Is it this one? http://spectrum.ieee.org/semiconductors/processors/transistor-aging

 

Lemme read and see if I remember anything.

I remember the guy talking about electrons or something in the transistors that go back and forth eventually get stuck or something.  

 

(Edit not the one)

[Marvzl1357]

Lemme read and see if I remember anything.

I remember the guy talking about electrons or something in the transistors that go back and forth eventually get stuck or something.  

 

(Edit not the one)

ok, nvm then

Well I guess greater heat=greater resistance and also thermal expansion, and so over time I guess the extra heat would probably physically alter the circuitry of the chip, thus causing it to lose its original performance.

Not sure though, everything just going off intuition

[Memories4K]

If anyone can help me find the article, I think I remember it being on Tom's hardware.  It may of been a post somewhere, I wish I would of bookmarked it. -.-

I am also very interested in this article as it would help me know how much i should push my future overclocks based on my cooling.

[Grrizz]

As said its both that affect its life span because both advance electron migration and in some cases thermal cycling etc. can be a factor.

As for power phases its a way of supplying more power with smaller/cheaper components whilst aiding in spreading heat output, each phase will have components like MOSFETs that can supply 1x its power rating constantly but something like 3x (arbitrary number) its power rating at a certain duty cycle, the down side to this is a lot more has to go into the circuit design because it brings in switching noise and timing issues.

[Devlafford]

Both.  The idea would be the more voltage you push, the safer it will be to keep it at a lower temperature.

There's an article somewhere I'll see if I can find it, it explains completely how chip degradation works, it's super interesting actually.

The jist of the idea would be, 1.4V 24/7 at 50C would be safer vs 1.35v @ 70c.

If anyone can help me find the article, I think I remember it being on Tom's hardware.  It may of been a post somewhere, I wish I would of bookmarked it. -.-

Thanks for answering that; if you could find that article, I would probably find that very interesting too. Thanks again! Gonna overclock this thing more, would help if temperatures on AMD weren't measured in "thermal margin"