fizzy said:
I also tried this last night after reading the above. Master card now on lower slot, slave/normal card on top slot. Changed card boot mode as mentioned and everything seems to work ok...
....but has not cleared up the interference issue. Mine is definately CPU usage related, as when I run a couple of copies of prime95 it becomes very noticeable - and thats not exactly gfx intensive.
Guess I had better start looking for some ferrite cores for the power cables...?
Update on issue - I raised an RMA with ocuk for the crossfire master GFX card. Thought I would do some further testing, to save any potential "no fault found" fees and such.
I used the GFX card in a single pcie slot nf4 based gigabyte board - works fine with no interference, even with the card being powered by the spare pcie power cable from the other (problem) machine (nice long leads on psu.
) as the 2nd pc's psu wasn't up to powering the card on its own. No intereference even when the main pc was running flat out too - so its not intereference down the pcie power connector...
...so did some more "messing" as it started to look more like a board issue. I managed to reduce the intereference considerably by moving my corsair dimms from slots A2 and B2 to A1 and B1. (wasn't sure if the dimm would fit under the overhanging fan on the heatsink)
Also noticed that the lower I put the dimm voltage, the more the intereference dissappeared. So I set it from "auto" to "2.6" and its a lot, lot better. Informed OcUK of this - and they raised a new RMA for me for the board, stating that "they had seen this before". I assume they are going to replace it..again...... hopefully third time lucky?
I guess this time its a problem with intereference/emi being created on the dimm power circuit/regulator, as I doubt that changing the vdimm has any effect on the cpu memory bus....
So - may get there eventually. Seriously putting me off getting an Asus board again though.