2x 7970's Still Black Screening

[Intention]

So the saga continues from here: https://bitcointalk.org/index.php?topic=84387
I ended up RMA'ing one of my cards due to it being faulty even just gaming on it's own or mining.

I got a replacement and it was okay for a little while...then later I found the main card ran full speed, the crossfire'd one would run at about 40% regardless of the clock settings.  About two weeks later I just constantly had the computer jump to a black screen with the displays going into hibernate with the system still on just not running Windows with the GPU fans going quiet but still running at this point.

Currently just going single card everything is all good... 980Clock/1030Mem @ 72C - 550MHash and I'll swap the card tomorrow and repeat the process but I need both of them going since well...that's why I spent so much on them.

I have an OCZ Z1000 PSU with the rail config like so: +3.3V@25A, +5V@25A, +12V1@83A, -12V@0.8A, +5VSB@6.0A

Using GPU-Z with 1 card only:
VDDC is 1.1V
MVDDC(no idea what this is): 0.006V
VDDC Current: 147A
VDDC Current in: 12.4 (avg)

Now I'm assuming VDDC Current In is the amps the card is drawing...I'm not sure where it's getting a current reading of 147 since the PSU itself should only be doing 83 on that rail.  Could my PSU be faulty or is there some treachery afoot?

[1715687236]

1715687236

[1715687236]

1715687236

[1715687236]

1715687236

[yunk3r]

My card was doing that. turned out it was the vrms overheating. my 7970 crashes when the vrm gets to be around 120 C. Given the reason this was happened was because I am water cooling without a full cover water block, but maybe your vrms run hot. try undervolting and see if that helps. Also download a program like HWinfo to monitor your vrm temps.

[Intention]

My card was doing that. turned out it was the vrms overheating. my 7970 crashes when the vrm gets to be around 120 C. Given the reason this was happened was because I am water cooling without a full cover water block, but maybe your vrms run hot. try undervolting and see if that helps. Also download a program like HWinfo to monitor your vrm temps.

Thanks I did give the program a try, right now VRM 1 is 80 VRM2 is 57.  For what it's worth I did have the clocks lower than 900Mhz for Core/Mem with both plugged in with the same voltage but I don't see how suddenly one would shoot to over 100 degrees.  I have an ROG board so I was using PCI-Express Slots 1 and 4 (optimal for 2 cards is 1&3) just to give them breathing room so it's not like they were wedged together.

[Plazzman]

the crossfire'd one would run at about 40% regardless of the clock settings.

Just a thought but have you tried with both cards and no crossfire?  Needle in the haystack approach...

[Intention]

the crossfire'd one would run at about 40% regardless of the clock settings.

Just a thought but have you tried with both cards and no crossfire?  Needle in the haystack approach...

Well this is what confuses me, it was doing it for the longest time I swapped to the second stock bios without the factory OC on it and it worked okay.

Regarding the VRM temps, the card features a back plate with a small cut out to them where the vrm's are but they weren't anything incredibly hot.  I'm using a HAF-X case so I was planning on sliding the fan that covers the GPU's back on to get some airflow blowing across the cards.  I was thinking of cutting up small thermal pads and getting little heatsinks to place on the backsides if it gets really bad - but I guess my point is it shouldn't even be like that running on reference clocks.

Edit:
24 hours of running VRM's are at 82C so I believe they are pretty safe.

[Intention]

So just a quick update.

I had my two 7970's clocked at reference and ran them for 20 hours using Diablo miner instead of CGMiner and had no dropping to a black screen.  I have absolutely no idea why this is the case but I appreciate the advice that was given.  I'll stick with Diablo and fiddle with CG for now if Diablo fails me then I'll be back.

[ar3]

in all honesty i have had two motherboards exhibit same kind of   issue.
total disclosure i have close to 15 7970s,all watercooled. 1 was a Rampage iii extreme with 4 cards,the other a maximus iv extreme with two cards

have you tried running your displays on the slave card and see if the issue changes.
can you try another mobo?


what mobo do you have?