Black Screen When Running Graphics-Intensive Games (GTX970M)

[mothreed]

I'm running an MSI GE62 with an NVIDIA GTX970M card on Build 10240 of Windows 10. The CPU is an Intel Core i7-6700-HQ @ 2.60 GHz and I'm running DDR4-2133, 16GB, for RAM.

As of the last few days, I've been unable to run most graphics-intensive games--on startup, my notebook goes to a black screen. Sometimes I can hear the audio still running, but I cannot CTRL+ALT+DEL or ALT+TAB out of the screen. The only thing that recovers it is a hard reset. I've had my notebook for about a year now and haven't had the issue before. I can also run relatively less intensive games without issue and have double-checked that both are running on the dedicated card.

Does anyone have any idea what the problem might be and how to fix it? I recently updated drivers for the card and integrated graphics, neither of which seemed to help.

Not sure if it's related, but I've had an issue recently as well where on boot the notebook would hang indefinitely at the MSI splash screen, forcing me to hard reset. That problem seems to happen less when I try to boot with the power cable and all USB devices unplugged. (I've updated drivers for all USB devices as well, on the chance that the boot issue and the black screen were related to something there, but no luck).

Any help at all would be appreciated. Thanks.

 

[Ralston18]

Open Event Viewer and look at the log entries.

May be confusing at first but search about in the different folders and look for red and yellow icons related to errors that have occurred.

Especially any that occur just before the black screen event. You can right-click an error for more details.

Look for an error code or some description with respect to the log entry.

May find multiple series of errors as some problem cascades into the black screen....

Google the error codes and look for a consensus before attempting corrective action. Then change only one thing at a time, keep notes so you can reverse as necessary.

Avoid any software or sites that offer to fix the problem - they will show up no matter what problem you are researching....

Check Event Viewer and post accordingly.

 

[mothreed]

So, the event viewer turned up a handful of things that seem to track with the black screens (either coincident with them, or occur just before or after). I'll just summarize what I found and did, since I'm really not sure what's relevant and what's benign or unrelated.

The obvious ones were critical ID 41 errors (sourced to kernel-power), which I'm pretty sure are just the hard resets after the black screens.

I also noticed some correlation with ID 3 errors (Kernel-EventTracing). I searched about, and it seems like the problem for others was that the sessions Readyboot and EventLog-AirSpaceChannel were exceeding the default max size for their stop conditions. So, I expanded the max size for both.

I'm also getting a truly massive number of ID 16385 errors (Security-SPP), about one per thirty seconds, which I gather is a problem with Task Scheduler. I checked and found that my Task Scheduler isn't working, and when I try to open it, it gives an infinitely repeating error telling me that there is an error and that it will try to re-start. Recommended fix was to change a registry value controlling the start of the scheduler, which I did (after manually backing up the registry and creating a restore point).

There were 1,652 instances of error 10005 (DistributedCOM) within five minutes on the day I started noticing the black screen issue, when I also had the hanging splash screen issue. Not sure

I've also had about 11 instances of error ID 1002 (Application Hang), which have only been happening since I've had problems, but which, on the other hand, don't seem to coincide with the black screens, so I think those are possibly unrelated.

There were also about 18 instances of error ID 1000 (Application Error) in the last three days, most of which coincide with the black screens. The faulting applications are various; the faulting modules are typically CortanaApi.dll or ntdll.dll. I'm really not sure what to do with these. I ran the System File Checker, which one source recommended. The checker turned up corrupt files, but was unable to fix them. I don't know enough about the process to be able to tell from the logfile exactly what those files are, so as of now, the ID 1000 errors are unresolved.

There were also a number of ID 5973 errors in the viewer noting failure of a number of apps to start, most related to Microsoft.Windows.Cortana.

Other errors that occurred around roughly the same time as the black screens, and which I either haven't figured out what to do about, or am wary of addressing yet, include numerous "Service Control Manager" errors (7001, 7031, and 7034 are the most numerous) and plenty of ID 37 warnings (Kernel-Processor-Power) noting reduced performance states for all processors.

Apologies for the rather scattershot response. Thus far, none of the changes I've made seen to have nipped the issue.

Open Event Viewer and look at the log entries.

May be confusing at first but search about in the different folders and look for red and yellow icons related to errors that have occurred.

Especially any that occur just before the black screen event. You can right-click an error for more details.

Look for an error code or some description with respect to the log entry.

May find multiple series of errors as some problem cascades into the black screen....

Google the error codes and look for a consensus before attempting corrective action. Then change only one thing at a time, keep notes so you can reverse as necessary.

Avoid any software or sites that offer to fix the problem - they will show up no matter what problem you are researching....

Check Event Viewer and post accordingly.

 

[Ralston18]

You have made a lot of progress.

Unfortunately all of those errors can be confusing and hard to untangle: cause vs effect, etc..

My suggestion is that you clear the logs and restart. Minimize the number of programs, services, processes being started as best you can.

Then you will have fewer entries/errors to sort out and sort through.

Restart and launch one of the graphics intensive games.

If the black screening stops then one of the startups may have been the culprit. (Or some combination perhaps...)

I am leaning towards a graphics card driver problem - especially if the black screens occur with all games.

However, I also am wondering about a possible PSU problem: What PSU do you have - rating, age, etc? Any signs of overheating, fan issues?

 

[mothreed]

Thanks for the advice and for getting back to me. I did the above and took down what the event viewer logged by the minute. I launched to a black screen again, so it looks like it's more than one of the startups (the only thing I left running on startup was Avast). This is what I logged:

----
Prior to program launch
----
- ID 16385 (Security-SPP)

----
4:50 (Program launched to black screen)
----
- 3 Errors: ID 7031 (Service Control Manager) [Contact Data_Session1, User Data Storage_Session1, User Data Access_Session1 terminated unexpectedly]

- 1 Warning: ID 10002 (WLAN-AutoConfig) [WLAN Extensibility Module has stopped]

----
4:51
----
- 1 Warning: ID 219 (Kernel-PnP) [The driver \Driver\WUDFRd failed to load for the device ACPI\ENE0110\5&3b0ec168&0.]

- 2 Warnings: ID 17 (WHEA-Logger) [Corrected hardware errors with PCI Express Root Port]

- 3 Errors: ID 1000 (Bonjour Service) [Two pertaining to mDNSCoreReceiveResponse, one noting that Local Hostname MSI.local is already in use, will try MSI-2.local

----
4:52 (From here until 4:54 am waiting to see if black screen changes)
----
- 8 Warnings: ID 37 (Kernel-Processor-Power) [limited by firmware]

- 1 warning: ID 4101 (Display) [Display driver igfx stopped responding and has been successfully recovered]


----
4:54 (Hard reset)
----
2 Errors: ID 6008 (EventLog, previous system shutdown was unexpected) and ID 1101 (Eventlog, audits events have been dropped by the transport)

1 Critical Error: ID 41 (Kernel-Power) [System reboot at black screen]

----
4:55
----
- 8 Warnings: ID 37 (Kernel-Processor-Power) [limited by firmware]

The PSU, if I understand correctly, is just the stock 150W adapter. It came with the notebook, so it's almost a year old exactly. No noticeable overheating (I always use a cooling pad--it does get rather warm without it, but I wouldn't say exceptionally so). The fan can be a little bit noisy sometimes, but not consistently or troublingly noisy.

You have made a lot of progress.

Unfortunately all of those errors can be confusing and hard to untangle: cause vs effect, etc..

My suggestion is that you clear the logs and restart. Minimize the number of programs, services, processes being started as best you can.

Then you will have fewer entries/errors to sort out and sort through.

Restart and launch one of the graphics intensive games.

If the black screening stops then one of the startups may have been the culprit. (Or some combination perhaps...)

I am leaning towards a graphics card driver problem - especially if the black screens occur with all games.

However, I also am wondering about a possible PSU problem: What PSU do you have - rating, age, etc? Any signs of overheating, fan issues?

 

[Ralston18]

Trusting that the PSU was/is correctly sized for the installed components along with some extra power margin I am leaning a bit more towards that PSU starting to fail.

Reason being that the problem is related to the heavy use of graphics and thus the power demand on the PSU.

And most of those logged errors/events are "shutdown, stopped, terminated" etc. are power related.

Being a stock PSU product means that the PSU is built in quantity as cheaply as possible with minimal attention to quality.

Do you have any warranty left? Try contacting the seller and report the problem - see what they have to say. Maybe check any related product forums and FAQs you can find. Look for similar problems being experienced by other MSI GE62 Apache owners/users.

Probably best to back up all data to another location - verify that the backups worked.

You could try a few more test rounds - similar to what you just did. Most likely will see more of the same - especially if I am correct regarding the PSU.

To be honest I am not sure what else to suggest right now. Hopefully someone else may spot something I am missing - no problem with that on my end.

 

[mothreed]

That's helpful--thanks for the insight. At this point, it's good to have anything at all to go on. I'll take a look. Out of curiosity, is there any chance that a PSU issue could also be responsible for black screens on boot/reset, or hangs on splash screen on startup? I've had plenty of those of late, and one thing I've noticed is that rebooting with the adapter unplugged sometimes gets me past the hangs.

Trusting that the PSU was/is correctly sized for the installed components along with some extra power margin I am leaning a bit more towards that PSU starting to fail.

Reason being that the problem is related to the heavy use of graphics and thus the power demand on the PSU.

And most of those logged errors/events are "shutdown, stopped, terminated" etc. are power related.

Being a stock PSU product means that the PSU is built in quantity as cheaply as possible with minimal attention to quality.

Do you have any warranty left? Try contacting the seller and report the problem - see what they have to say. Maybe check any related product forums and FAQs you can find. Look for similar problems being experienced by other MSI GE62 Apache owners/users.

Probably best to back up all data to another location - verify that the backups worked.

You could try a few more test rounds - similar to what you just did. Most likely will see more of the same - especially if I am correct regarding the PSU.

To be honest I am not sure what else to suggest right now. Hopefully someone else may spot something I am missing - no problem with that on my end.

 

[Ralston18]

My general sense is that PSU problems can manifest in any number of ways and not always consistent with respect to the device in question.

And sans adapter means less load and a failure is simply forestalled a bit. Any component that does not get the power it needs will falter at best and then some mixed (as you have seen) cascade of events occur ending with the black screen.