I have a Commando on its way to me. I see many tests in this thread but IMHO these tests have been anything indicative of stability. A minimum of an 8hr Orthos blend should be mandatory and 12hrs is even better (I've had Orthos error out 9hr 29min 9sec). I've also found that Orthos should be at a priority of 3 instead of the default of 1 and at the same time the tester should use his PC for normal apps.
I guess if you're testing for benchie numbers only then a 1hr run is fine. But those who claim stability with such tests do others a disfavor.
Even Orthos is not the be-all end-all of bench testing. My brother's ICFX3200 can run a 24hr Orthos blend yet it locks when running one of his intensive Oracle scripts after 15min. My advice: find your max stable overclock with Orthos then bump down 100MHz to be safe. That's what he did and he hasn't had any problems at all.
I have done testing with an eVGA 680i, Gigabyte DS3, Infinity 975, and DFI ICFX3200. So far my max oc was with the ICFX3200 at 3651MHz. I hope to achieve more with the Commando.
One question: Has anyone experienced lockups with a USB mouse with 0803? I've heard it can be an issue.