Conclusion
Since Core 2 Duo is currently dropping from the upper class into the mainstream, a reasonable platform seemed to be the best choice for this article. The MSI board we used comes with an overclocking switch, allowing users to easily set their bus speed from FSB1066 to FSB1333 or FSB1600. We used the BIOS settings and different speed increments to overclock a Core 2 Duo E8600 to as much as 4.26 GHz.
The focus of this article is efficiency, hence it was more important to determine the best clock speed with a focus on performance per watt. The P45D3 Neo-F was helpful for this article idea because the three-phase voltage regulator is typically more efficient than the multi-phase circuits found on enthusiast motherboards. The downside is limited capabilities for overclocking a Socket LGA 775 quad-core to speeds of 4 GHz, due to possible power limitations. Still, users looking for quad-core power should probably wait for Intel’s P55 platform and Clarkdale processors, which are expected later this year.
3.9 GHz Wins!
Although the 4.26 GHz setting provided the most performance, the 3.9 GHz overclock turns out to be the winner of this efficiency analysis. The clock speed increase provides a nice speed bump when compared to the 3.33 GHz stock speed without bumping power consumption up to excessive levels. If you are looking to get the best performance paired with the lowest effective power consumption, the fastest possible speed without a voltage increase should be your goal.
We found it more impressive to see that the performance per watt during a PCMark Vantage run was better than the default Core 2 Duo E8600. This means that all overclocked settings provided better efficiency than the 3.33 GHz stock speed. Clearly, Intel’s Wolfdale core could be running at 3.66 or probably slightly faster speeds without being less efficient! Unfortunately, it is pretty safe to say that such speeds won’t be introduced—they don’t fit into Intel’s multi-core strategy anymore—although an E8700 model at 3.5 GHz might be coming soon.
