Comparing Different Aspects of Overclocking
In order to compare the advantages of overclocking, I've broken the results down into four different performance metrics: stock as offered by the motherboard, an overclock of the memory only, an overclock of the CPU only, and a set of numbers with both CPU and memory overclocked I call "Full OC". This will give you a chance to see how only overclocking the CPU and using lower-performance memory works, and what using the maximum divider offered by the platform natively implies. But I will go over some power consumption numbers first.
|Load Condition||CPU Voltage||Ring Voltage||Idle Power||Load Power|
|Stock Clocks||1.041 V||1.148 V||22W||81W|
|Memory-Only OC||1.041 V||1.148 V||26W||85W|
|CPU-Only OC||1.296 V||1.148 V||35W||129W|
|Full OC||1.296 V||1.148 V||36W||133W|
Power consumption at these different intervals is quite interesting. Going from all cores at 3900 MHz, as defined by the motherboard, to 4600 MHz, as defined by my maximum overclock settings, is only a 700 MHz increase in frequency, or a bit under 20%, but power consumption increased by about 50%. Pushing much further with this chip is clearly not possible since the power draw is not linear to the clock increase and boosted voltage requirements. Also interesting is that the increased memory clocks only boosted power consumption by 4 W with both stock and overclocked CPU speeds. I actually expected slightly higher values when overclocked, but that was definitely not the case. I feel that most users, no matter what clock they end up with, will probably see similar power consumption figures before running into a cooling wall, which also reflects the information given by OEM overclocking guides; so while I might take a slightly different approach here, the similarities cannot be ignored.