"YES! IT'S ALIVE!"

Finally, 3.6 GHz stable! (i stopped at 7h 14 mins no errors)
Now i'll try to lower voltages, seems that the tricky part was, guess what, finding the right
GTL ref values, wich are kinda way off the charts, but who cares, and for the first time adjusting the value for the core that failed actually resulted in a stable system!
A big THX to the guys at DFI, as i would've never come up with those values on my own, and i only needed to up one core by 2 points to get the system from 2h stable to 7h stable!