@Dopamin3, I think you missed my last post, which was made just before you posted yours, but in case if you didn't allow me to enlighten you on the situation a bit more:
Many people have told me things about the NB/Chipset voltage, and not everybody seemed to have the same exact answer for what it did really. Based on my own overclocking experiences, I came up with the theory that the NB/Chipset voltage directly affects the reference clock (FSB) and nothing else.
I decided to test this on a NB OC of 2750MHz at the CPU/NB voltage of 1.275 (if I recall correctly) because I knew the OC wasn't stable at those settings (nor at any other CPU/NB voltage below 1.4v, based on previous experiences anyway). So what I did was I slowly increased the NB voltage instead of the CPU/NB, because if my theory was right, then no matter what the NB voltage, the NB OC would be just as unstable as last time. I proved my theory wrong, however.
Every time I did increase the NB voltage, I got a bit further in Prime95 time and time again. It eventually got to the point where one or two cores would fail like 30 minutes in, and the others would continue on with no problems. After I got to 1.415v, I decided to finally increase the CPU/NB voltage. When I got to 1.35v, I decided it was stable enough to try this OC along with a mild 3.5GHz CPU OC.
When I booted into Windows, I immediately went to Prime95 and started running the test, but one core couldn't even start because the fft was too large or something. At that point, I knew something wasn't right, and that's when I found out that only one stick was working.
Hopefully you understand what I was doing and why. I clearly was not upping random voltages - I knew what they were, and I was putting what everyone and I had thought about those voltages to the test. Unfortunately, my RAM died in the middle of the process.
With that out of the way, what RAM do you guys recommend I get for OC'ing now? Edited by TheRockMonsi - 12/4/10 at 2:19am