• Notice regarding image hosts and attachments. See here.

  • Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Crossfire 7950 keeps crashing PC

So ULPS was disabled in the registry already so it's not that.

Did the driver sweeper thing then rebooted and installed beta drivers/CAPS and rebooted again, same problem, so it's not that.

Checked GPU-Z and found that card 1 was 16X PCIE as expected, card 2 was 4X PCIE as expected (limitation of the mobo), so it's not that either.

I have a monitor attached to each GPU at the moment and crossfire unplugged/disabled, is there any way I can bench both cards at once, one per GPU? I would like to stress them both out in this way to see if adding load to a second card pushes it over the limits. I tried Furmark but it always uses GPU0 as it's renderer.

I don't think it's overheating as such as even when temps don't break above 60 it can still crash. I even downclocked both GPU's by several hundred Mhz on core and mem, and set the fans to 100% on both and it still crashed very quickly (though perhaps a little less quickly than normal).

After it crashes both cards are still relatively cool to the touch.
 
Try it with only 1 stick or RAM in and see what happens and check both sticks. I think it was LtMatt that mentioned in another post that RAM issues become more apparent when using xfire.
 
With one stick in the same problem occurs. Will try the other stick later. One thing that is interesting is that I can't complete many memory passes on the windows memory check. Seems to get hung up some runs in and just sits there until I press escape to exit the app.

Could be nothing but don't have any spare disks to burn memtest onto to test them with that.
 
Update: I went ahead and got a new PSU - a Be Quiet Dark Power Pro 850W http://www.overclockers.co.uk/showproduct.php?prodid=CA-064-BQ&groupid=701&catid=123&subcat=1083 - the behaviour is still the same so it wasn't my PSU at fault...

I may keep this PSU anyway it seems much quieter and better than my old Corsair 1000W job was.

Any other suggestions welcome. I will do a BIOS update next on the motherboard but I'm not ruling out a problem with the motherboard or card overheating issue (though overheating at 60 degrees celcius seems too low to me...).

The board I have doesn't have much room between the two cards, they are almost touching at the fans, as the fan on the top card is only slightly above the bottom of the bottom card.

I may even see if Windows 7 has the same issue, could be something funky with Windows 8 :(
 
Sadly I did that already, the only thing I haven't really done is tried with Win 7 instead of Win 8 so if it's still doing it after this BIOS update that will be my next step.
 
With one stick in the same problem occurs. Will try the other stick later. One thing that is interesting is that I can't complete many memory passes on the windows memory check. Seems to get hung up some runs in and just sits there until I press escape to exit the app.

Could be nothing but don't have any spare disks to burn memtest onto to test them with that.

Have you managed to do a Memtest run yet? Be interesting to see the results. I'm running xfire fine in Win 8, so doubt it's that.
 
I'm running memtest86+ now on one of the sticks after creating a bootable usb stick so will let you know what it does. Any other suggestions welcome.

The odd thing I find is that it's gpu2 that seems to go high on temps according to msi afterburner. Gpu1 should get hotter really as it's the top card and has less airflow. Gpu1 seems fine though where gpu2 quickly gets hotter and seems to cause a crash.
 
Ok memtest86+ ran on the stick of memory (the single stick in my PC at the moment) through 3 passes successfully, took about 3 hours, no errors. I don't think it's the memory.

Running out of culprits now. I'd really like a way to stress both cards individually on different screens to see if I can trigger a problem without crossfire enabled, anyone got any suggestions?
 
So I figured this out in the end.

I took the card out of the top pci-e 16x slot totally and only attached one of the cards to the bottom pci-e 4x slot.

As soon as I try and stress the card in this slot at all it very quickly crashes the whole system!

So individually the cards work fine (tested both in top slot) but the bottom pci-e slot seems broken...

Time for a motherboard replacement I think. Not sure I'll get the same model again after this may look to get a better one which is more reliable.
 
Glad you managed to get to the bottom of it. The board you have is a decent one, and it will be replaced under an RMA, so might be worth sticking with it.
 
Back
Top Bottom