Of course, those readers who are interested in the new generation Intel processors would hardly be satisfied with Kentsfield screenshots only. Everyone wanted to see some real overclocking results and the reports about some corresponding experiments have already hit the pages of XtremeSystems.org forums.
Before we start investigating the overclocking potential of this CPU let's take a closer look at one very interesting peculiarity of this processor. Just like its dual-core predecessor it allows disabling ?excessive?cores. Although in this case three cores out of four will be disabled. As you know, Conroe has shared L2 cache that is still available as a whole even if one of the cores has been disabled. Kentsfield consists of two Conroe dies, each featuring two cores and 4MB of L2 cache. By the way, the CPU-Z utility reflects L2 cache as 4MB, while it simply doesn't see the second 4MB at all :)
So, if you disable three cores of the Kentsfield processor, then the remaining core will only work with 4MB L2 cache that remained in the active half of the CPU.
Maybe there will be a way to disable only two cores of the Kentsfield CPU. However, there will hardly be the need for that, mostly just to increase the overclocking potential.
With the nominal Vcore, the Kentfield processor overclocked from 2.4GHz to 3.2GHz with the air cooler. It is quite nice for a quad-core CPU, I should say. C1 core stepping sample should be able to go farther than that with higher Vcore settings and more powerful cooling solution.
By the way, they didn't increase the multiplier of the Kentsfield processor during their overclocking experiments. Actually, all Extreme series CPUs should have unlocked multipliers. Kentsfield should also have one. Maybe the CPU owner simply didn't check out this feature