OC'd for over a year

[RangerXLT8]

Originally Posted by Gig-O-Ram

At one point, I tried to increase from 2.76ghz to an even 2.8ghz with my voltage set on 1.456v (according to CPUZ, or 1.53vcore according to nView Monitor), and my system became unstable, especially when gaming. So, I took it back to 2.76 and everything was fine. I didn't want to push my luck.

And I just did a major cleanup of dust from the entire system a few weeks back, which was about the first really good cleaning it had since I put it together. I'm gonna have to be more diligent in the future, since my chipset was running pretty hot most of the time (lots of dust in it), and my Big T was pretty dirty too. But the chipset is doing much better now, and hopefully the AS5 I put under it is helping some.

My stock voltage is 1.3v if I remember correctly, with a stock speed of 2.2Ghz. I was never sure if 38% overclocked would be considered significant, considering that I've seen some people acheive nearly 100% OC's (with some pretty over-the-top cooling, mind you!). What is electron migration and what can it do?

Is 3.46Ghz your overclock on your 920?

Electron migration is like the corrosion caused by a flowing river.

Increased voltage restructures the path that current flows through the CPU.

3.46GHZ is where this 920 runs stable now... 920 Default freq is 2.8GHz/

[Gig-O-Ram]

Originally Posted by RangerXLT8

Electron migration is like the corrosion caused by a flowing river.

Increased voltage restructures the path that current flows through the CPU.

3.46GHZ is where this 920 runs stable now... 920 Default freq is 2.8GHz/

So, that said, it sounds like the CPU pins or the socket itself would show signs of what would look like burn damage as a result of the higher voltage. But when I recently took my system apart to clean it and install a new controller, I had removed the HSF to apply new AS5, and when I removed the CPU and looked for damage, I saw no signs of burning anywhere - unless the damage would be internal.

[gvblake22]

Ranger is talking about changes to the silicon substrate and transistors at the microscopic and atomic level, nothing you would be able to ever see with the naked eye.

[screwballl]

This is also why many cpus after being run for a year or more with no overclocking, if you try to overclock it it will not take much more than a 5MHz OC.

Get the OC started early and keep it at the same speed for long periods of time, changing the speed once a month is more likely to kill your cpu

[Gig-O-Ram]

Originally Posted by screwballl

This is also why many cpus after being run for a year or more with no overclocking, if you try to overclock it it will not take much more than a 5MHz OC.

Get the OC started early and keep it at the same speed for long periods of time, changing the speed once a month is more likely to kill your cpu

Sensitive little things, aren't they? It sounds like the CPU gets cured for either way to run it...Either overclock it or don't and then leave it that way. I've heard of curing TIM, but I had no idea CPUs were set up like that.