Intel’s Core i7-990X, -980X, and -2600K are all fairly unique in the company’s lineup for their unlocked multipliers. The feature’s a boon to overclocking the two Gulftown-based chips, since you don’t have to screw around with the 133 MHz BCLK. However, it’s an absolute necessity on the Sandy Bridge-based processors because you can’t screw around with its 100 MHz BCLK.
Here’s the thing, though. Intel bundles its Extreme Edition parts with a big, beefy DBX-B thermal solution for enhanced overclocking. But it includes a much dinkier heatsink/fan combo with the K-series SKUs.
Using Intel’s retail DBX-B, I was able to push the Core i7-990X to 4.4 GHz at stock voltage.
Using the same cooler, I pushed the Core i7-980X to 4.26 GHz at its stock voltage.
Armed with the cooler bundled with Intel’s Core i7-2600K, I was barely able to push 4 GHz at the chip’s stock voltage before bouncing off 80 degrees Celsius (I don’t like operating in excess of 80 degrees for extended periods of time). Swapping out for a Thermalright MUX-120, I was able to sit comfortably at 4.3 GHz at stock voltage. Any faster than that and the Core i7-2600K would crash, needing more voltage, which you could comfortably give it, since the hottest core was only hovering around 60 degrees. In essence, if you want to overclock a K-series chip with any real seriousness, add $50 or $60 to your budget for more capable aftermarket cooling.
We of course know that all three processors could go faster—but their relative performance won’t change much. The real purpose of overclocking at all was simply to show that both Gulftown processors and the Sandy Bridge CPU have headroom in them, and the conclusions drawn at stock frequencies apply here, too. In the applications where the -990X was faster, it remains faster. Anything that prefers the -2600K’s more efficient architecture continues to.
|Core i7-990X||Core i7-980X||Core i7-2600K|
|3ds Max 2010||1:45||1:48||2:12|
|Adobe Photoshop CS5||:57||:59||1:15|
|Adobe Premiere Pro CS5||12:15||12:15||16:40|
|ABBYY FineReader 10||1:05||1:06||1:19|