Originally Posted by radaja
ih didnt notice you were on a P45.with a quad core and 4GB you might need 1.24v to 1.32v mch for 400fsb.i know for 4GB and Q9650 i needed 1.44v mch
1.24v didn't work, so I've jumped it to 1.3v
And it's gotten 7 errors in under a minute. o_O
Now it's gone to 11 in under 90 seconds. Test 7 only.
Now 17 in under 2 minutes
Originally Posted by kremtok
I found that there's an upper limit to the effectiveness of raising northbridge voltage, and it looks like that's what you're seeing too. Keep it where it ran best, then try raising the memory voltage just a tad.
Do you have some other RAM you can try out in the board? Do you have another board where you can try the RAM? Everything I see points to a motherboard problem, specifically with running your memory in dual channel mode. That's the problem we need to sort out.
1.2v is where it ran for the longest, and I've got it running now at that. 3m into the test, no errors so far which is much better than 1.22, 1.24 and 1.3. That makes no sense though, why would increasing the voltage cause more instability?
I don't have any more RAM or another DDR3 board. I really can't afford this to be a problem with the motherboard. I bought it from a US ebay store MWave (I think) and Gigabyte don't offer warranty for this board in Europe.
So far, there's not a single component that I've not had a problem with in this build - first the original XFX motherboards, then the graphics card, then the CPU and now the RAM or Motherboard again. Edited by [Adz] - 11/1/10 at 3:52pm