Intel Core i7-2600K (and friends) Sandy Bridge Processor Review
Overclocking on Sandy Bridge
Overclocking the 2nd Gen Intel Core Processor family is quite a bit different than even the Lynnfield processors released in 2009. The changes that were implemented with the next-generation Turbo Boost technology are to blame/thank for these changes as we are no longer able to simply push and push against a thermal wall by adding more voltage and better cooling. Because the new CPUs are able to monitor its current thermal state much more accurately (necessary for that new Turbo tech) Intel has implemented safety precautions that prevent the previously used overclocking methods, or at least makes them much more difficult to implement and utilize.
For some of you, this will be a disappointment - the days of the LN2 wars could very well be behind us. (Hopefully more innovative overclockers than I will figure something out though.) But in my opinion, overclocking has gotten easier and less stressful - if you have a K-series CPU that is. Increasing the frequency of your processor via the base clock (which is 100 MHz on these) is severely limited with changes of only about 109 MHz being possible in my testing, at ANY multiplier. That is a drastic change from how we were able to push the Lynnfield/Bloomfield processors in the past.
The primary way to overclock the new SNB processors in my testing has been to adjust the Turbo multipliers to staggeringly high rates. Without adjusting the voltages in any way, using the ASUS P8P67 Deluxe motherboard, I was able to set the CPU to a 45x maximum Turbo multiplier that brought the CPU's frequency to 4.5 GHz. Even better, because the C-stepping and other power management features can be left on while doing this type of overclocking, the CPU ran at the same 1600 MHz at idle and only ran at faster speeds when applications and tests were being run.
According to what I am seeing and discussions with a few motherboard vendors, there do seem to be more limits on the overclocking potential of the new SNB CPUs in general - you will likely find a frequency where you simply can't climb past with any amount of cooling or voltage increases. And that limit will likely be able to be reached on modest air cooling and modest (if any) voltage increases. This is what I mean by overclocking being "less stressful" - there seems to be much less of the process:
- Set a speed
- Reboot, see if it crashes
While we can't be 100% sure this will always be the case, it would appear that users can feel confident they are getting the most out of the CPU with much less work than they had to do previously. In other words, SNB might close the gap between what you can do and what the overclocking elite can do.
That could change of course, and we'd be disappointed in companies like ASUS and MSI didn't continue to push the limits of these processors and find new ways to enable those LN2-type folks to really push things further once again.
Once we return from CES we plan on spending a lot more time with our sample processors (and few more we ordered) and a handful of motherboard to see what else we can find!