Unknown Issue - Possibly related to OC =/

[Mark617]

System -

http://www.newegg.com/Product/Product.aspx?Item=N82E16820231314

http://www.newegg.com/Product/Product.aspx?Item=N82E16813128503&Tpk=gigabyte z68mx b3

OCZ agility 3 ssd

windows 7 home prem

Every so often, usually between 6 hours and 3 days I will get a shut down. It only occurs when playing everquest 2, so obviously this could be an issue. But it is not like I am in windows that long at a time either so it is hard to use that as a reference point. Event viewer really does not show anything helpful before the crash, just the power issue after (from me shutting it down).

What happens is as follows - the game itself freezes, but vent in the background is still working and audible, although my keyboard does not work. I can hear others and my mouse still moves and when i alt tab i revert back to the windows mouse instead of the image of the game mouse, but can't do anything (num lock caps lock dont work). Eventually the computer locks up, but in and of itself it does not shut down. I have to reset it or power it off manually.

The system is stable in prime 95 at 4.3 ghz and stock voltages with the only settings changed as follows - power states all disabled ie c1 eist etc. p95 was stable for about 26 hours then i turned it off. Ran blend tests. memtest86+ is stable for around 3 hours and I turned it off.

** One weird thing, if I oc then memtest shows a lower memory speed because it slows my bclck down. IE at 3.3 ghz it shows 1600 speed at 100 bclck but at say, 4.5 ghz it shows around 1176 speed at 73 bclk. Is this a usual change?

Anyway, I am going to run it at 3.3 ghz and play and see if I can mimic the issue thereby narrowing it down. But, that being said, I wanted to get advice / opinions here as soon as possible and i will edit shortly. off to work ~

Oh and anyone with this MB if they can help with settings i would appreciate it. I will try and post the options soon to get some advice.

thanks as always

edit - to add a few things which may or may not be relevant.

Heat isnt an issue. its WC and in an ~70 degree F room real temp never shows above 59 F.

one time I THINK i hit shut down and the computer did a restart. i cant be positive though, but hitting shut down and having a restrart would lead me to think a possible MB issue. this would have been with the f2 bios, it has not reoccured with f5.

i use the latest radeon 6950 driver but dont use control center. never seemed to have a need.

It just feels like a memory, GPU or HD issue. I have nothing to back that up but when i have had CPU OC issues in the past they have been full halts, whereas this seems to be more a failure somewhere else in the chain. again, i have nothing to back this up other than past events and a feeling.

I suppose my biggest hurdle now is to figure out what programs / utilities i can use to narrow down the problem piece of hardware / software. I dont have any spare hardware to switch out for testing =(

I was thinking of lowering the mem speed to 1333 from 1600 and messing with the timings. The only ram voltage option i have is dram voltage and its running at stock of 1.5v. Speaking of voltages, no matter what i do, cpu-z tells me the vcore is 1.068v. In the BIOS if i change vcore it shows a change in the bios but not cpu-z. like i said though, for now vcore is on auto. Why is this? an issue between my mb and the cpu-z program?

Sorry everything is not really organized, I am at work throwing everything i can think of out there to help with troubleshooting. im usually quite OCD about my grammar and sentence / paragraph structure lol.

I will update as stress tests are done in sequence for 24+ hrs over the next week if i cant solve it / find help sooner here ty

 

[Djinni]

First steps are as you noted to reset all settings back to stock values: Remove the OC on your CPU, RAM, GPU.

Then I would check that your GPU is seated and powered properly, (and that it's getting enough power) since you pointed out that the game freezes and doesn't accept input, while background programs continue to run.

What happens is as follows - the game itself freezes, but vent in the background is still working and audible, although my keyboard does not work. I can hear others and my mouse still moves and when i alt tab i revert back to the windows mouse instead of the image of the game mouse, but can't do anything (num lock caps lock dont work). Eventually the computer locks up, but in and of itself it does not shut down. I have to reset it or power it off manually.

 

[Mark617]

How do I go about checkng the power to the GPU? I mean besides checking it's physically connected

This is the PSU I use. I know its older but its been reliable and seemed to have the proper connections to do the job:

http://www.newegg.com/Product/Product.aspx?Item=N82E16817703009

* I am guessing something like MSI afterburner coupled with 3dmark and watching temps / voltages, ya? Or whatever program is compatible with my radeon card.

EDIT - In addition, within the BIOS are settings for GPU core (1.110 v) but I was under the impression this is for the on board. The option for on board is set to no PCI-E exists - which states the on board only works when a pci-e card is not deteced. The 6950 is the only card and sits in the pci-e x16 slot.

Then I would check that your GPU is seated and powered properly, (and that it's getting enough power) since you pointed out that the game freezes and doesn't accept input, while background programs continue to run.

If it was not seated properly and not connected via the 2 pci power connectors would I even get a post / boot? The issue, if it is power, seems to be an overload which I would think points to a bad PSU or not enough power available for the gpu, no?

 

[Conumdrum]

Try it again with no overclocks. If it's good it's probably a OC setting. I have heard that AMD likes ram at 1333 and then timings set tighter.

The PSU looks okay for now. It's got enough juice.

 

[Mark617]

Yep - running on default now and if things go well I am going to OC, but drop the mem down to 1333 and see how it goes.

Now, if I was to OC but leave the voltages the same, would there be additional draw for power being that the cpu is OC'd? I am assuming no, since I did not change the voltages, correct?

The reason I ask is if it came to it, I wonder if the possible (previous) over volt + OC could have caused an instability outside of the OC. Meaning, the PSU was barely providing the power at stock settings and the excess draw caused the system to fail - ie power, not the OC being the issue.

Anyway, that last question is more for future reference if and when things fail.

 

[Mark617]

Computer just froze on stock settings. Dropped memory down to 1333 to see if it helps.

Cant figure this out now as it actually seems to be getting worse as time goes on / settings lowered.

I am beginning to think this is more related to the gpu (which is not / has not been OC'd), power, or the software. Best way based on sticky above to stress gpu still one of the 3d marks run on loop and monitor temps through CCC / another program?

EDIT - what 3d mark is best? would it be better to buy the 3d mark 11 to run it or try and find the older ones and simply run the nature tests? how important is the dx 11 to the actual stressing portion of the computer?

Edit #2 - 3d mark 11 ran for over 4 hours, no issues. Ran the custom version, 1680x1050, priority real time, looped and simply let all the tests and demos run. **memory was dropped down to 1333 though for stability in game and I did not bother to turn it back up to 1600 **

Edit #3 - 9 hours memtest86+ passed with no erros. I will test it longer tonight but needed to do some work on the computer this morning.