Originally Posted by bratas
Wow can you contradict yourself anymore?
If indeed the data sheet is fact and default voltage needed to run the cpu at stock is 1.520. Then 1.520 with an OC is safe. Now if I saw a chip running at 1.520 for stock I'd be returning that thing so dam fast it wouldn't be funny.
People have been running their OC for months at 1.52 with no problems as of yet.
With that argument, you're BOTH contradicting yourselves.
First of all, the VID of the CPU is based on Intel's own calibration, on what the CPU's need to run stable at their default and turbo multipliers, with the STOCK heatsink and fan, at up to the tjunction point before thermal throttling begins (~98C).
Two CPU's can have completely different VIDs, and also can give off different heat, even at the exact same set vcore.
No CPU will EVER have a default VID of 1.52v at stock voltages. Ever.
The VID is basically calibrated so that the CPU can be fully operational at stock and guaranteed turbo multipliers (e.g. 34-38 on 2600k), and also that the CPU can be functional at higher multipliers. The default VID for the chip will increase based on its current multiplier. The CPU should be able to at least complete its power on self test and make it into the BIOS, with the default VID (provided you aren't at a multiplier wall).
Now, what happens when you start getting to high multipliers means that stability is no longer guaranteed. For example, if a CPU has a VID of 1.39v at 5 ghz, and you didn't pass the MP wall, you should be able to POST into the BIOS at 1.39v. That doesn't mean that the CPU will be STABLE if you put a load on it. But the BIOS itself should run.
Also, the presence of vdroop throws a monkey wrench into the works; while the CPU may be stable at 1.39v in the BIOS, if vdroop drops the vcore down to 1.30v at 5 ghz, the CPU might not even complete POST.
So yes, the VID is based on VRD 12, and is calibrated for each multiplier, but ONLY guaranteed for standard operating speeds, e.g. 3.3 to 3.9 ghz. But the CPU can still request a higher VID--UP to 1.52v, at higher multipliers.
Intel knows that people overclock, and they expect that people who will put that vcore into the CPU, will have sufficient cooling to handle it.
Also, high vcore increases the chances of degradation, of course, but this does NOT mean that low vcore PREVENTS it completey. There have been 2500k's that were able to load windows at 1.29v and 5 ghz, without stress testing, on the very first boot, then needed 1.31v on the very next boot, then 1.33 about an hour later. Then stabilized at 1.35v without more degradation.
So yes, it varies from CPU to CPU. Virtually ALL cpu's will degrade SLIGHTLY--also known as "break in", from the very first boot you have them on, to the very next boot, at their maximum speeds, with the lowest possible vcores. This "break-in" happens to everyone. But people don't usually know what their ABSOLUTE LOWEST possible vcore for stability is, at a high overclock right away, so they will possibly have a little degradation without knowing it, as the speed they're running at is already unstable anyway...