Artifacts and Crashing; pretty sure it's not the GPU

[jhames]

MOBO: MSI X99s Gaming 9
CPU: Intel i7-5820k
CPU Cooler: Corsair H110 w/ 4 Phanteks 140mm Fans in Push-Pull Config
GPU: MSI GTX 970 4G (x2)
RAM: Corsair Vengeance 16GB DDR4-2400
PSU: EVGA SuperNova G2 1000w
STORAGE1: Plextor 256GB M.2 SATA
STORAGE2: WD Red 1TB
STORAGE3: WD Blue 1TB
CHASSIS: Phanteks Enthoo Luxe
Monitor 1: ASUS ROG Swift
Monitor 2: BenQ GL2760H
Monitor 3: BenQ GL2760H

So I've been having problems with my desktop lately with artifacts and crashing. I'm pretty sure it's NOT the GPU too. About two months ago, my computer would get artifacts and would crash whenever I was in SLI. I ruled it out to be my second graphic card so I RMA'd it. I got it back a few days ago and it still caused a crash. I tried it by itself and immediately crashed my system when I got to the desktop. When I removed it, my first graphic card started to give me problems, but after replugging my wires in (one of the DVIs were loose), they seemed to go away. Yesterday the artifacts/crashing came back full force and haven't gone away since.

At this point, I ordered a brand new graphic card (exact same one) and put it into my system by itself and... immediate artifacts when getting to the desktop with a crash about 10 seconds later. So I think I've ruled out my graphic cards. The chances of having 3-4 bad ones in a row is extraordinarily unlikely. I'm not exactly sure what my next steps should be so I have a few questions:

(1) Which of the above components could cause the above problems?

My immediate thoughts would be the PSU, but typically wouldn't that cause the whole system not to power on?
RAM is a possibility. I'll try removing a stick (4 sticks total) to see if that fixes it.
Motherboard seems unlikely.
I doubt it's the CPU.

(2) What is the best way to test a PSU? I have a multimeter, but I'm not exactly sure how to approach how to use it to test the PSU.

(3) Could it be a/the monitor(s) that are causing it? I've never experienced it personally, but I thought I've read somewhere that monitors could cause issues like this.

The computer appears to completely freeze with no sound, responsiveness, etc.

Any thoughts, suggestions, comments, w/e will be greatly appreciated. I'm getting rather desperate as the release of The Witcher 3 nears. 🙁

 

[i7Baby]

Download HWInfo and run Sensors. Game for 10 mins. Post the Sensors results - usually 3 screenshots in TinyPic or similar. To see how your hardware is running.


Your stuff looks fine but who knows? Something's not right.

 

[jhames]

It shouldn't be a hardware incompatibility. The system ran has been running fine for months.

I actually already have HWINFO installed. I'm just not able to even access the desktop for more than 2 minutes before the whole computer freezes and I have to unplug it.

Could a failing PSU cause these symptoms?

 

[little_me]

yes, Failing PSU could do it.
in SLI, where you before first experienced problems, both cards draw power, out of sli, other is usually just mostly idle. This would of course depend a bit on where monitors are connected and so on.

Bios might or might not allow you to check the voltages that the PSU gives although while in Bios, it isn't really under normal load which hwinfo is better for.
If you have spare PSU, trying with that is worth a shot in my opinion. If spare has less oomph, trying with only one GPU might also be worth it

 

[i7Baby]

Test the PSU per Step 23 of http://www.tomshardware.com/forum/261145-31-perform-steps-posting-post-boot-video-problems and post the volateg readings you get.

 

[jhames]

Okay, I'll check the voltages on my PSU later and post back. Assuming I ruled out the GPU(s), if it's not the PSU, the only other thing that could cause this is the motherboard, correct?

 

[jhames]

I haven't gotten around to checking the voltages quite yet, but I troubleshooted some other components:

Chassis: pretty much impossible to cause this.
RAM: Tried each card separately in as well as in different slots so it's unlikely to be that.
Motherboard: I tried putting the graphics card in a different PCIE slot and the artifacts and crash still appeared. I don't know if other aspects of the MOBO are broken, but it appears not to be the PCIE slot.
PSU: Haven't checked the voltages yet; will post when I do.
GPU: I think it ruled it out when I tried a total of four GPUs.
CPU: Haven't tested. To be honest, I'm not even sure how to. I don't believe this would be the cause though.
Monitor: Tried each of my monitors individually; artifacts and crashing occured on all of them.
Monitor Cables: Tried DVI, Displayport, and HDMI. Artifacts and crashing occur for each.
Storage: Haven't checked, but highly improbable it would be that.

Software Issue: This is where it gets interesting. I noticed that the artifacts and crashing occur when I get to the desktop. At the login screen (I'm running Windows 8.1 x64 if it's relevant), nothing occurs. I left it at the screen for over any hour occassionally moving the mouse to see if anything would happen then and nothing.

Booting into another user account (I have an ArcGIS server account) and the system loaded just fine. Booting into safe mode also worked just fine. In safe mode, I uninstalled the graphics driver and booted back into the standard OS and everything appeared to work just fine UNTIL nvidia control panel decided to reinstall itself (I'm pretty sure it's part of my chipset drivers, so when it's uninstalled, it tends to reinstall itself). After it reinstalls itself, the artifacts immediately occur and the system freezes/crashes.

With that being said and done, this leads me to believe it's one of three things:

A software issue; I would like another opinion before I do a format and clean install (I have tons of software to install so it'll take a day to redo everything

The PSU; could it be that when the nvidia drivers are installed, the card fully "activates" and overstresses the PSU, causing artifacts and crashing? I'm not familar with how a software driver causes a computer to interact with the hardware.

The GPU(s); I still think it's highly unlikely since I've tried 4 different GPUs, but hey, weird stuff happens.

Thoughts?

 

[i7Baby]

Post the voltages

 

[jhames]

Here is the HWINFO log from a different user account. That fact that I'm able to log into a secondary account (created in safe mode) and operate the computer with no problems makes it seem like its definitely a software issue. I even put both graphic cards in the computer. The new user account also has nvidia drivers installed. Something on my account is causing the computer to crash. I'm currently backing up some miscellaneous files now from the secondary account so I haven't been able to take readings from the PSU itself. I'm going to run a few stress tests while on the secondary account to see if it's stable. I'll get the readings with the multimeter after I run the stress tests and finish backing stuff up. Thanks for the help so far mate.

 

[i7Baby]

Your primary GPU isn't firing up.

 

[jhames]

My monitor(s) are plugged into that GPU so it's definitely on, I'll run kombuster real quick and see if it fire ups.

 

[i7Baby]

You can see the difference in GPU clocks in the 3rd panel of Sensors.

 

[jhames]

That's actually my secondary GPU. For some reason, I have them labelled backwards in HWINFO. 😛 My secondary GPU was disabled because I forgot to enable SLI when I was adding and removing my GPUs to test if they were dead. Here is a link a screenshot from HWINFO with SLI enabled.


EDIT: Temps and everything read fine after performing Furmark for 15 minutes with SLI enabled. I just booted into my normal account and the system immediately crashes. I'm going to rule it as a software issue and perform a complete format and reinstall. If that doesn't work, I'll post back!