XFX Radeon HD 7970 Black Edition - System Crash problems

[spikernum1]

I'm an experienced custom PC builder, and I was fortunate enough to build a pretty good PC last month.

My Problem:
My computer will randomly "shut off" regardless of the type of use/load it is put under.
I could be playing BF3 for 6 hours, and have no problem with it. Start browsing some web pages, and it will power off. That being said, I have had instances where I will crash during BF3 or other heavy loads. I just want to point out that it isn't isolated to any single situation. It is seemingly completely random.

My Troubleshooting:
I have run IBT and Prime95 to make sure my CPU and RAM are working fine - they can go maxed for over 24 hours.
I have run FurMark and get my video card up to 78C without it crashing.
I can run 3DMark without it crashing.
I have run Memtest86+ and confirmed my RAM is fine.
I have undone my CPU overclock

Other Notes:
When the computer "shuts off", the fans will still be going, the 7970 fan will actually ramp up for about 3 seconds before coming back down to it's regular low speed. Also, my G15 keyboard lights and my mouse laser will still be on - so USB power is still coming to them.
I have the option to restart on blue screen turned OFF, but I don't get a blue screen (not one that I can see anyways...)
I also do not see ANYTHING in the Event Viewer. All I see is the event in which I have to power down the system to get out of the crash state.
My monitors (one 27" LED Samsung using HDMI, one 22" LCD Samsung using DVI) will go into "no signal" state.
CTRL+ALT+DEL will not do anything.
The only thing that brings it out of this state is doing a hard reset or just pressing the power button.

Specs:
GPU: XFX Radeon HD 7970 OC Black Edition 1000MHZ 3GB 5.7GHZ GDDR5 (GPU is at stock OC speed that the Black Edition uses)
PSU: OCZ ZX Series 850W ATX12V 80Plus Gold 70A
CPU: Intel Core i5 2500K 3.3GHz (OC to 4.0GHz)
Cool: Corsair H100
RAM: Corsair Vengeance 8GB 2x4GB DDR3-1600 CL9-9-9-24 Dual Channel (x2 = 16GB Total)
Mobo: ASRock Z68 Extreme3 Gen3 ATX LGA1155
SSD: AData 120GB 6Gb/s Sata III

Drivers/Updates:
I have the very latest Windows Updates. Latest drivers and firmware for all my hardware.
Catalyst version 12.3

Interesting Note:
As I mentioned above, I use two monitors. However, the crashes seem to completely stop if I only use my primary monitor (27" LED Samsung over HDMI) and do not have my other monitor plugged in. I haven't swapped the two to see if it is actually a monitor driver issue... but I think it is less likely than a faulty graphics card or graphics driver issue.

Suspicion:
I suspect a faulty graphics card or fault graphics driver. My first thought is to simply return the XFX card and get a different brand. Although I have never had problems with XFX in the past, I've heard many horror stories - and this may be my first experience of one.


Please let me know if I can do anything else to troubleshoot that I'm not thinking of, or if you also have experienced something similar.

I've been dealing with this for about a month now, and it is frustrating not knowing whether you should keep such an expensive piece of hardware if it doesn't work properly.

*Edit: Modified subject.

 

[rolli59]

If the secondary monitor works fine as single, I would RMA the card!

 

[spikernum1]

Update:
I had one monitor going for almost the entire day without any crashes, until just now. Got the same problem.

This time, I was able to CTRL+ALT+DEL to restart my computer instead of doing a hard reset. When the computer booted and I hit the Windows Login screen (I heard it through speakers), my monitor was still not finding any source. I actually logged into Windows blind, and then powered off the system. After powering back on - the video game back on.

Still no logs of any errors in the Event Viewer... Only the fact that it restarted without cleanly shutting down first.

IMO This pretty much narrows it down to PSU, Mobo, or GPU.

Since the easiest thing to replace on the system is the GPU, I figure I should just RMA that and see if a new one resolves my issues.

I am a little frightened that it could be the PSU or Mobo.... because I don't want to have to rebuild from scratch again and do all that cable mgmt again. lazy me...

I'm hoping someone can pipe in with "yeah it is definitely your GPU" to make me feel better

But seriously, do the symptoms fit the PSU or Mobo being defective moreso than the GPU?

 

[rolli59]

That is the problem nobody can say for sure! GPU can though be eliminated with trying it in another computer or try another GPU in the faulty rig.

 

[spikernum1]

Okay, one more thing I'm curious about.

http://i40.tinypic.com/24ctamq.png

It shows that my CPU Turbo is going to clock at 6GHz... I hope that isn't the issue here.

 

[rolli59]

Okay, one more thing I'm curious about.

http://i40.tinypic.com/24ctamq.png

It shows that my CPU Turbo is going to clock at 6GHz... I hope that isn't the issue here.

That would certainly crash it!

 

[spikernum1]

It doesn't make sense though... I undid my OC and turned off Intel SpeedBoost and it still displayed the CPU Turbo as 6000MHz.

I have a feeling this can't be accurate.

When my CPU is under full load, it stops at my max setting (3.3GHz stock, or 4.0GHz OC)


I'm more convinced it is the video card than the CPU. Since if the CPU would cause the crash, the computer would simply restart immediately - and my monitors wouldn't STAY off during the next startup.

 

[spikernum1]

I'm trying one more stability test before I RMA the video card.

Used Trixx to turn off ULPS and bump up the clock slightly.

I hope it isn't just faulty ULPS combined with drivers causing this trouble...

 

[spikernum1]

Well, the issue is still happening.

I'm going to RMA the card this evening and get a new card and run the same tests.

 

[spikernum1]

UPDATE:
I tested using my old 9800 GTX, and the problem continued to happen.

Then, using the 9800 still, tried swapping PCI-E slots to see if perhaps the slot was faulty, and it started crashing even MORE! I was able to recreate this one quite easily... just power on, wait about 3 minutes, and the system would begin to slow down to a crawl, then eventually blue screen. This is different than what I experienced with the 7970... where THAT would just instant restart and I would lose monitor power.

I booted up in Safe Mode to determine whether it might be a driver problem, and I still got a blue screen after a few minutes of it just sitting idle.

Swapped the 9800 back to the original slot has resolved the blue screen issue. And it hasn't crashed yet... but it crashed in this slot originally.


Ideas?!?!?!?!

 

[MrCommunistGen]

Do you have a spare PSU of sufficient wattage you can "splice in"? In testing PSUs on systems with heavy cable management I've set up a second PSU on top of the case (or wherever the cables reach), unhooked the PSU inside the case from all components and plugged in the "external" PSU in its place. I'd suggest trying that.

 

[spikernum1]

Turns out it was two things. Video card caused the screens to lose power - RMA'd and got a new Gigabyte 7970. Doesn't happen anymore.

But my SSD was also defective - which caused the blue screens and other random issues.

Thankfully it was under warranty, so I got it replaced and fresh installed everything.

All is working fine now

 

[rolli59]

Great that you are sorted, can be hard enough to find one faulty part let alone two!

 

[JACK45]

Hi there !

Simple to fix it, I have had the same problems with my 2 XFX 7970 BE with 3 Monitors @ 5760x1080 res.

Just make sure you have all Updates done (Win7/8 drivers etc..) and then download "MSI AFTERBURNER", install it go to "Settings" to the Fan Control and enable it ! Configure it and play around with the Settings to stay all the time under 81°C !
Don't forget to enable in the common adjustments the "Start with Windows" button and finally "Apply overclocking at system startup" button on the main screen of the Afterburner. That’s it!!

Test all with MSI KOMBUSTOR!
And then play your Games and have fun ) !

Jack!