So I started work on this last night... method 2 gave me a little trouble at first... looking at the ASUS SPD table in the Tools area showed me different timings then what the xmp profile had. After a few resarts I figured that out... then on to the voltage changes you mentioned... setting the SA, aI/O and dI/O via offsets has yielded some promising results.
One of two things would happen... crash or BSOD would occur just as windows was loading up, sometimes it would get as far as the log in screen but usually during the colored floating balls. I would incrementally increase the voltage and try again. QUESTION: Since I didn't know any better I have been increasing all three of these roughly the same percentage each time i needed to increase them. Should I continue this or is there a better way to approach it? My current plan is that once I get it passing benchmarks/stresses to back off each one individually in order to fine tune... if i see behavior i don't like after a decrement, I know to go back up one and then move to the next voltage (does that make sense?)
So it got late last night and I ended up leaving off what I think will end up allowing me a stable realbench pass. (should be able to update tonight with final numbers (fingers crossed). My realbench passes started off crashing during the second phase but now they are consistently crashing during the heavy multitasking portion, I managed to get through that test last night a couple times with no problem before calling it a night. I'll start tonight by attempting to run a full pass or three and then move on to occt.
Now... If i understand correctly... the System Agent voltage is what is fed to memory controller, pcie controller and the onboard graphics controller? and the I/O voltages are for the interaction between mobo, ram, pcie, etc. ? is this (layman's) accurate?
Asus Maximus Formula VI | 4670K | Corsair Vengeance (4x4) 2400mhz | Corsair H100i | Cooler Master HAF-X | Corsair AX860 | GTX760 | Samsung 840 120 SSD | ASUS VG248QE | Corsair K70 | Corsair M95 |