• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

quick question about what kills cpu/gpu in overclocking

xerocool

Senior member
ok, so i know people can fool around with the multiplier, fsb speeds, voltages, and probably alot more. my question is which one of these if any or all kill the lifetime or fry a cpu or gpu.

basically i'm looking at two scenarios- immediate damage and long-term damage for each modifiable aspect of overclocking.

i.e. if i increase the fsb speeds to a point where my system can't handle it, does that do any immediate damage to my computer or does reseting it to a lesser fsb basically make it like it had never happened? also if i let it run at high fsb speeds for long periods of time, does that decrease the lifetime of my equipment?
 
top 3 reasons (in no particular order)

1) voltage
2) running at unstable speeds for long periods
3) heat

running at unstable speeds can cause the chip to "burn down" to a point that it will no longer run at that speed at all. once its "burned down", not even adding voltage will help it.

its one of the reasons why i (and other AT members) preach "stability, stability, stability" 😉

HTH 🙂
 
>>>i know people can fool around with the multiplier, fsb speeds, voltages, and probably alot more. my question is which one of these if any or all kill the lifetime or fry a cpu or gpu

The most likely culprit in a fried CPU is a melted trace somewhere in the chip (i.e. one of the vias that carries current through the circuitry). By far, the most likely way to make this happen is raising the voltage, as doing so also increases the amount of current. I suspect you're unlikely to get the circuitry hot enough to melt merely by running the chip faster (i.e. modified multiplier or FSB). 70C may be pretty hot to us humans, but to the materials used in a silicon chip, it's really no big deal.

My suggestion - play with anything you like - EXCEPT the voltage.
 
ElectroMigration. Over the natural life span of a CPU, only a finite number on electrons can travel from point A to Point B. When this happens, the CPU cannot operate any longer or at least not normally. It may just look like the CPU is burned out but its really that the transistors cant move anything like when it was new. This is a laymans explanation because I dont know the exact terminology for this stuff.

So if you increase voltage, run it a unstabe speeds for a long long time, or if the heat gets to it, are all contributing to this migration in a quicker fashion thus shortening the life of the CPU. Might take a few years life off of the CPU. But if you are an enthusiast, you will never own a CPU long enough to see this happen because your always looking to get the next best thing constantly.

Results will vary from one CPU to the next as usual.

Voltage I would think would be the most instantaneous damaging factor over anything else. Heat would be the next. P4's clock themselves down when temperatures reach unsafe levels. I am not sure if AMD has implemented this feature as I am not an AMD user.

Keys
 
Back
Top