Originally Posted by munaim1
Not sure about the different versions of prime but I will look into it
Yeah, it could be something on my end, hard to tell. For one though the 26.6 version has 1344FFT and the 25.11 does not. If you force it to run that FFT it acts a bit weird.
Also Tunagoblin (with the same mobo as me) said at some point he needed to enable spread spectrum, somewhat randomly/unexpectedly. He also stated that he needed to raise
CPU PLL to get the best results and he ended up with 1.832 (read this on xtremesystems).
I've been doing a lot of testing today and taking notes and the first thing I noticed is that raising CPU PLL appears to be better for me (I tried lowering it several times with bad results). I say 'appears' because man there are just too many variables at this point, and 1344 seems really random and touchy. Anyway, I have ended up at 1.832 after several tests and the results seem to be better as a whole.
The next thing I tried was enabling Spread Spectrum. I know, I know.... it's supposed to be enabled in the first place. I had it disabled since the beginning, and since my overclocking was going quite well I felt no reason to change it back. On this board with SS enabled the BCLK changes from 100 to 99.8 so that was another reason. Anyway I tested 4-5 times with it enabled (and got the BCLK to come out to 100.1) and things seem pretty good right now.
I also played with VTT. When I raised it too 1.129 I got, for the first time ever, a core failure / rounding error (during 1344 after like 1 minute flat). Wasn't sure what to make of it so I lowered it back to 1.103. Nothing below 1.103 was conclusive either way.
Anyway I'll let you guys know what ends up working for me. Hoping to get this blend test done once the AC is fixed here, ugh its hot.