Windows 10: Reboot playing Overwatch (or gaming), or running FurMark -- No BSOD

For roughly the last two months, I've been experiencing reboots when playing Overwatch, or sometimes when playing other games (e.g., Metal Gear Solid: TPP, The Witcher 3). Sometimes, these reboots won't happen immediately when loading the game, and the computer may allow me to play for an hour or more until it reboots. Once it does, it will usually reboot my computer as soon as I load into the menu of the game (i.e., once the first reboot of the day starts, my system will reliably crash if I try to play Overwatch if I try launching into the game in consecutive attempts). Sometimes, leaving my computer running for a while before trying to play the game will allow me to actually play for another extended amount of time until this cycle repeats, which I find to be quite odd.

There's no BSOD, the Event Viewer doesn't seem to have useful information (symptoms rather than causes), and it doesn't seem like dump files are being created.

I've tried reinstalling some drivers (though not exhaustively): first for my NVIDIA GeForce GTX 970 driver, rolling back to a previous driver with no success, and drivers for my Asus motherboard. I've also reinstalled Windows with no improvements, and have recently reaffirmed that the image is healthy through DISM in the command line.

I also went down a deep rabbit-hole with an infamous AMDA00 driver issue with ASUS motherboards, but I no longer believe this to be the actual cause of the reboots.

Assuming it might not be a software problem, I've also looked into physical issues. I've checked the power cables, and they seem to be properly snug onto the PSU and GPU. I've stress-tested the CPU with Prime95, and results and temperatures seemed pretty normal (maxed out at 60 C). I've used MemTest86 for the RAM -- 8 passes went by with success. I've also checked the integrity of my SSD and HDD that I use for Windows -- both returned with no errors.

This leads me to believe it's an issue with the PSU or the GPU. The first time I ran FurMark, I ran it for a couple of minutes before stopping it and it seemed okay just as a test run. However, I have not been able to run FurMark since then without it instantly rebooting my computer (similar behavior to when gaming). However, I am very skeptical that the GPU is faulty -- temperatures when gaming never exceed 75 C, and FurMark reboots the computer when the GPU would only be at it's idle temperature (I believe around 30-40 C). I've used a tool to test the VRAM, and that also returned with no errors.

My current PSU is the EVGA SuperNOVA NEX650G.

I'm also attaching the ZIP file generated by this forum's BSOD diagnostic tool.

One thing to note: I started dual-booting with Ubuntu on a separate SSD within the month I started experiencing these problems, but reset the BIOS settings to the defaults and have also disconnected the Ubuntu drive. I wouldn't be surprised if this was the cause of this instability, but I don't know how it could be related.

If I could get some more pointers about how to debug this, or reassurance that my PSU could be faulty, I'd greatly appreciate it. I'm hoping this is a resolvable software-side issue, but I'm prepared to get new hardware if need be. Thank you!

:)

 

Desktop PC crashes and shuts down after playing some games for a few minutes.

How demanding the game is does not matter, several high demand games run just fine, the only ones that I know that crash are listed in the OP, Overwatch and Doom are not light games. Overwatch actualy gets the temps higher than any of the other games but
is one I play all the time and never get any crashes.

Just typing that I realize that should be alarm bells, but I would rather play Overwatch most of the time anyway.

 

Sudden artifacts, freezing and BSODs under some stress

Since yesterday, my laptop seems to be malfunctioning. While playing games like Battlefield4 and Overwatch, after some time I get artifacts and my screen freezes(game keeps running and windows is still active, can still hear notifications and game sounds).
After this I can exit the game through task manager and relaunch the game, untill all this happens again.

During this recurring problem, I also got the occasional BSOD: INTERNAL_VIDEO_SCHEDULER_ERROR

This problem seemed to start out of nowhere. It also seems not to be temperature related(GPU reaches only 65C and CPU 75C).

Things I have tried but didn't work:

-Reinstalling gpu driver.

-Installing older gpu driver.

Things I have tried that seemed to work:

-Set TDRdelay to 8 in registry.

The last thing seemed to work for a while, untill again while playing Overwatch, I saw artifacts, the screen froze and got a BSOD with the message THREAD_EXCEPTION_NOT_HANDLED.

Here is a link to the last 3 DMP files, MSinfo and speccy:

DMP,MSinfo,Specs.zip

 

If the power supply was the problem I would expect the PC to just shut down and not restart by itself.

If it is software related then you can follow the clean boot troubleshooting method to disable various software programs during start-up.

Other things to check include:

• Examine all fans and heatsinks for dust build up and clean them thoroughly.
• Make sure all connectors and cards are fully inserted in the Motherboard.
• Check all your drives for errors using a utility like HD Tune.

The inability to run Furmark is likely to be a clue. Can you try another GPU for testing purposes?

 

Hey Phil, thanks for getting back to me. I'll try the clean boot troubleshooting method now and post back to this thread with any more findings I get.

To address your other points:

• I've examined the fans and heatsinks for dust, and cleared them out. Didn't seem to have an effect.
• I still need to check all the connectors and cards for the motherboard.
• HD Tune seems to have found no issues in either my OS SSD or storage HDD.

I'll be able to use another GPU, but it would take about a week to have access to it (when my roommate's parts come in for his PC build).

I didn't know that the restart would be indicative that the PSU isn't the issue ... interesting. How come?

 

If the PSU is failing and cuts out it should not restart itself. However, if it is giving an unstable voltage supply I can see that it might trigger the motherboard/OS to cause a restart, you should be able to see voltage instability with a monitor program like speccy or HWmonitor.

 

So I noticed a motherboard cable was slightly loose. I readjusted it, ran FurMark fine for about 35 minutes (max GPU temperature of 80 C). Stopped FurMark, and tried another run -- my computer rebooted before I even saw the pop-up window of the stress test.
As for my voltages on the motherboard and GPU: using HWMonitor, it seemed like the voltages were pretty stable/consistent on the GPU, keeping at around 0.981 V when I successfully ran FurMark just now. The motherboard had a bit more fluctuation, but nothign that seemed too outrageous (fluctuations of about 0.01 V in a couple of areas, but mainly +12V and VIN4. Here's a screenshot of HWMonitor for context when I was running FurMark.

Attachment 126298

Going to triple-check cables, then do safe boot debugging. Do the voltages seem reasonable?

 

So I've booted into clean boot, and am still exhibiting similar behavior. I may be able to run FurMark once, perhaps even twice, but then booting it up another time will trigger a reboot. After the first reboot, any further attempts instantly trigger a reboot as well.

UPDATE: I borrowed my other roommate's graphics card, an EVGA GeForce GTX 1070, and the same issue occurred. Doesn't seem like a GPU issue.

 

The GPU temperatures are getting to the top end of what I would be comfortable with. Are you overclocking your system in any way?

 

No, I haven't over clocked my GPA myself at all.

 

Hey Phil -- I caved and ordered a PSU, the same one I have now. After swapping it out, I was able to run FurMark consecutively without any issues, which I was not able to do with the old PSU.

If I continue to not experience any issues for the next couple of days, I'll mark this thread as solved, unless you have anymore comments. If not, thank you so much for your help -- greatly appreciated to have someone knowledgeable chime in personally on my issue! This has definitely made me much more capable of debugging Windows issues now. *Biggrin

 

Great news! You should check the new voltages and compare with the old one. Be interesting to see if you can spot a difference.

 

I ran FurMark again with HWMonitor on, and got the following results (posted below as a screenshot).

I exited FurMark as I was about to hit the predicted peak of 80 C, just cause I didn't want to keep the card running at that temperature. I could've sworn I heard a single beep come directly from my computer around the time I neared 80 C, which caught me off guard -- a bit scary, but I'm hoping it's non-consequential.

Only difference in voltages that catches my eye is that the motherboard +12V reading reached a minimum of 12.0 V, whereas it previously dipped to 11.808 V on my old PSU. The other readings on the motherboard and GPU don't seem strange.

Attachment 126431