My only other suggestion is to turn LLC (Load Line Calibration) on as it reduces the vdroop you experience when the CPU is put under heavy load. You should be able to lower your Vcore in the bios.

I know what you mean in regards to running 4 less threads than the CPU is able to provide. It was the reason I sold my 940 for the current 920 D0 I run now. In all honestly, having HT on at 4 GHz hasn't made my PC "feel" any quicker, but it was just something I wanted to do.

You should be able to hit 4.00GHz and drop your vcore considerably with HT off, maybe to around 1.275V - 1.300V in the bios (with LLC enabled) depending on your chip.