Here's what I am planning on doing with mine when I build it in the next couple of days.
Step 1: I am going to drop all multipliers and test the boards max BCLK. Then I will test the systems max memory, QPI, PLL, core, and Uncore clocks at stock voltage. They I will measure the effects of adding voltage to each (temp, stability, additional oc headroom). Then I will back everything off about 20% and find the minimum stable voltage for each. Then I will back off another 20% and take another reading.
Eventually, I will end up with several graphs that will show voltage vs frequency and voltage vs temperature. I will then use the graphs to find the optimal overclock and voltages. I will then take the derivatives of the graphs and interpolate a mean coefficient Harmonic frequency using my bat computer.
Step 2: after all that I will spend 3 days scratching my head at why it's not stable. Then suddenly I will get phone call by a mysterious fellow telling me that he's captured the mayor's daughter and she's slowly descending into a pit of acid. I will then use my Pentium 90 and GPS modded TI 83 calculator to triangulate the origin of the call within the city but due to a rounding error I ended up at a Taco Bell instead of the fellow's hideout. Taking advantage of a disadvantageous situation... I order a crunchwrap.
Step 3: ???
Step 4: OMG 8GHZ OC ON AIR!!!!!!!!
That's the plan anyways....
PS: the 1st paragraph is true, and I might also make some graphs also but unfortunately I don't have a bat computer so I can't interpolate a mean coefficient Harmonic frequency... perhaps the i7 could do it for me, but could I trust the results at 8ghz????