Overclocking headroom question

[Anteaus]

I don't have any direct experience with overclocking, but I've been reading alot about it and I became curious about something. I noticed that when people talk about overclocking some chips, in this case various Intel i7 chips, there seems to be a great deal of OC potential in out of the box. I've seen claims where some chips have been OC as much as 0.5-1.0 Ghz above stock speeds with "relative" ease (stock cooling) and in some cases hitting 1.5-2 Ghz above stock in extreme cases. I'm speaking in generalizations, since obviously there will be a disparity from chip to chip as to what it can be OC to and remain stable.

Where I'm confused is to why this overhead exists? It seems like a waste to market these chips in a way that grossly underestimates there capabilities. It just seems academic to me that if I had a group of CPUs I would want to market them at the highest possible stable speed of the group, with possible overclocking potential of the "better" chips somewhere in the 5-10 percent range.

It seems like the vast majority of these intel chips are being overclocked from 25-35 percent over the rated speed, and that seems a bit odd to me, as if Intel doesn't know how to property bin CPUs anymore.

I'm sure there is probably a technical reason for all this and I'm probably just missing a piece to the puzzle, but would someone please explain it to me?

 

[T_Yamamoto]

Some chips can over clock differently. Also, companies want to have this headroom so that their chips won't blow up essentially

 

[Belegost]

Temperature, power, variability, market segmentation.

1st to hit the 1Ghz+overclock range demands good cooling and power, and will generally overrun standard tdp limits for most OEMs. Second, within those limits AMD/Intel have to guarantee that all chips will function %100 stable, that encourages conservative clock speeds, especially since a string of bad chips have stability issues could mar reputation a lot.

That holds down the top end clock speeds, everything else gets binned to segment the market for specific price points.

And on Intel's side, lack of motivation. AMD no longer provides competition at the high speed parts, so why not be conservative. Even better SB and forward provide such a controlled overclocking environment that Intel can charge extra to those customers willing to push the limits, note how the K series has a sizable minimum cost for both the chip and platform. Sounds like gravy to me...

 

[Flipped Gazelle]

First of all, this OC overhead has existed for a long time. I overclocked my Motorola 68000 from 8 mhz to 10 mhz. On the PC front, I began with overclocking a 386/33 to 40mhz.

Among other reasons, like product differentiation, the headroom exists because the mfr (Intel, in this case) wants to ensure it's CPU's can within a specific thermal envelope. Once you start overclocking, power consumption and heat can increase dramatically.

 

[Idontcare]

Power consumption goes through the roof is the bottom line.