Question Troubleshooting help! Potentially Faulty Graphics Drivers That Won't Fix?

[Jergister]

Okay so here's the deal, I have been plagued by constant crashes while gaming for over a year and I have tried everything under the sun to try and troubleshoot it. I have recently replaced my motherboard and now have been getting blue screens that have indicated that the errors are associated with graphics drivers. I have dump files for the most recent crash but I'm not sure how to post them. I have tried reinstalling drivers multiple times as well as I have recently completely wiped my computer and reinstalled windows to no avail. I did notice that when I had a fresh copy I somehow acquired the most recent NVIDIA drivers before I had even downloaded the software. These crashes usually happen 5-10 minutes after I've opened up a game and I've gotten either a blue screen or I've also had the game simply close before as well.

Here are my specs: Intel I9 9900KF, NVIDIA gtx 1070Ti, STRIX Z390-H gaming MOBO, and 32gb DDR4 RAM and I'm using a NZXT H510 case.

Anything that can help is greatly appreciated.

 

[Ralston18]

What game or games crash? What game or games, if any, do not crash?

PSU: make, model, wattage, age, condition (original to build, new, refurbished, used)?

Look in Reliability History and Event Viewer for error codes, warnings, and even informational events that occur just before or at the time of the crashes.

How are the graphics drivers being installed?

Are you using some third party installer or directly downloading the drivers via the applicable manufacturer's website?

Use direct downloads and then manually install and configure. No third party tools or installers.

 

[Jergister]

What game or games crash? What game or games, if any, do not crash?

PSU: make, model, wattage, age, condition (original to build, new, refurbished, used)?

Look in Reliability History and Event Viewer for error codes, warnings, and even informational events that occur just before or at the time of the crashes.

How are the graphics drivers being installed?

Are you using some third party installer or directly downloading the drivers via the applicable manufacturer's website?

Use direct downloads and then manually install and configure. No third party tools or installers.

Before I upgraded my motherboard, only GTA, Sea of Thieves, and Dying Light were the ones that crashed, but without any blue screen. After I changed my motherboard every game started crashing, games like CS:GO, Rocket League, League of Legends that wouldn't crash before are now crashing.

PSU is EVGA 750W, used by my friend and had no problems, 3-4 years old. I checked the voltages on it with nothing appearing strange.

Reliability monitor is showing some stuff, but I'm not exactly sure how to get the data I need from event viewer.

 

[Jergister]

What game or games crash? What game or games, if any, do not crash?

PSU: make, model, wattage, age, condition (original to build, new, refurbished, used)?

Look in Reliability History and Event Viewer for error codes, warnings, and even informational events that occur just before or at the time of the crashes.

How are the graphics drivers being installed?

Are you using some third party installer or directly downloading the drivers via the applicable manufacturer's website?

Use direct downloads and then manually install and configure. No third party tools or installers.

Just tried with a manual install of drivers and got a crash again.

 

[Ralston18]

Reliability History is much more user friendly and can be very revealing with regards to errors and patterns.

Event Viewer is much more difficult to follow, navigate about, and understand. Mostly you have to just stick with it for awhile to get a sense of things and then dig out the data you need.

Here is an older link from within this Forum:

https://forums.tomshardware.com/faq/how-to-use-windows-10-event-viewer.2752289/

And another tutorial:

https://www.windowscentral.com/how-use-event-viewer-windows-10

Overall, if between Reliability History and Event Viewer, you are seeing increasing numbers of errors and varying errors that makes the PSU suspect.

Especially being 3 - 4 years old and likely used for heavy gaming. The PSU may be nearing its' designed in EOL (End of Life) and starting to falter and fail.

Most likely at times of peak power demands.

Try swapping in a known working PSU if possible. If modular, use only the modular cables that come with the PSU.

 

[Jergister]

Reliability History is much more user friendly and can be very revealing with regards to errors and patterns.

Event Viewer is much more difficult to follow, navigate about, and understand. Mostly you have to just stick with it for awhile to get a sense of things and then dig out the data you need.

Here is an older link from within this Forum:

https://forums.tomshardware.com/faq/how-to-use-windows-10-event-viewer.2752289/

And another tutorial:

https://www.windowscentral.com/how-use-event-viewer-windows-10

Overall, if between Reliability History and Event Viewer, you are seeing increasing numbers of errors and varying errors that makes the PSU suspect.

Especially being 3 - 4 years old and likely used for heavy gaming. The PSU may be nearing its' designed in EOL (End of Life) and starting to falter and fail.

Most likely at times of peak power demands.

Try swapping in a known working PSU if possible. If modular, use only the modular cables that come with the PSU.

I can look into that, I'm gonna have to buy one since I don't have anything here right now. Would a failing power supply give codes that usually indicate corrupt drivers?

 

[Ralston18]

"Would a failing power supply give codes that usually indicate corrupt drivers? "

Likely not but then again problems can generate errors codes that may not be directly related to the problem. The other part of that is that many error codes, etc. tend to be too general/broad to be truly helpful. Akin to the "Check Engine" light.

For example, a power glitch to a GPU might disrupt some code being executed and Windows flags that disruption as corrupted driver code.

And it all gets messier in that PSU's provide three different voltages (3.3, 5, and 12) to varying components. So one thing may work or appear to work while something does not. The end results being even more confusion with respect to error codes.

The pattern I look for is increasing numbers of varying errors. Improper shutdowns included. ( Sometimes forced on the end user in order to get the system to reboot.)

Very much becomes a matter of elimination: Manually download, reinstall, and reconfigure drivers. May or may not help.

Ensuring that all components, cables, RAM, jumpers, and case connections are fully and firmly in place. A loose connection can wreak havoc. Subsequent power losses corrupt files.

Windows has tools to fix its' own files: "sfc /scannow" and "dism" for example. Those tools may or may not work. Then you may need a clean reinstall to recover.

Sometimes the culprit is correctly and quickly discovered and fixed thereafter. Other times - not so much.

 

[Jergister]

"Would a failing power supply give codes that usually indicate corrupt drivers? "

Likely not but then again problems can generate errors codes that may not be directly related to the problem. The other part of that is that many error codes, etc. tend to be too general/broad to be truly helpful. Akin to the "Check Engine" light.

For example, a power glitch to a GPU might disrupt some code being executed and Windows flags that disruption as corrupted driver code.

And it all gets messier in that PSU's provide three different voltages (3.3, 5, and 12) to varying components. So one thing may work or appear to work while something does not. The end results being even more confusion with respect to error codes.

The pattern I look for is increasing numbers of varying errors. Improper shutdowns included. ( Sometimes forced on the end user in order to get the system to reboot.)

Very much becomes a matter of elimination: Manually download, reinstall, and reconfigure drivers. May or may not help.

Ensuring that all components, cables, RAM, jumpers, and case connections are fully and firmly in place. A loose connection can wreak havoc. Subsequent power losses corrupt files.

Windows has tools to fix its' own files: "sfc /scannow" and "dism" for example. Those tools may or may not work. Then you may need a clean reinstall to recover.

Sometimes the culprit is correctly and quickly discovered and fixed thereafter. Other times - not so much.

Gotcha, its gonna be a minute before I can get the power supply, but I'll give that a shot and see if anything comes of it.

 

[Jergister]

"Would a failing power supply give codes that usually indicate corrupt drivers? "

Likely not but then again problems can generate errors codes that may not be directly related to the problem. The other part of that is that many error codes, etc. tend to be too general/broad to be truly helpful. Akin to the "Check Engine" light.

For example, a power glitch to a GPU might disrupt some code being executed and Windows flags that disruption as corrupted driver code.

And it all gets messier in that PSU's provide three different voltages (3.3, 5, and 12) to varying components. So one thing may work or appear to work while something does not. The end results being even more confusion with respect to error codes.

The pattern I look for is increasing numbers of varying errors. Improper shutdowns included. ( Sometimes forced on the end user in order to get the system to reboot.)

Very much becomes a matter of elimination: Manually download, reinstall, and reconfigure drivers. May or may not help.

Ensuring that all components, cables, RAM, jumpers, and case connections are fully and firmly in place. A loose connection can wreak havoc. Subsequent power losses corrupt files.

Windows has tools to fix its' own files: "sfc /scannow" and "dism" for example. Those tools may or may not work. Then you may need a clean reinstall to recover.

Sometimes the culprit is correctly and quickly discovered and fixed thereafter. Other times - not so much.

Took a while but finally got a power supply installed. Ran good for a little bit but just got another blue screen with the code system_service_exception. Looking at the minidump it looks like
svchost.exe is what crashed maybe, but I'm not sure what that could mean.

 

[Ralston18]

Look in Reliability History and Event Viewer.

What error codes, warnings, or even informational events - if any - preceded the blue screen?

 

[Jergister]

Look in Reliability History and Event Viewer.

What error codes, warnings, or even informational events - if any - preceded the blue screen?

It happened while I was playing CSGO and prior to the blue screen the game just randomly closed twice with no codes or anything.

 

[Ralston18]

Power down, unplug, open the case.

Clean out dust and debris.

Double check by sight and feel that all cards, connectors, RAM, jumpers, and case connections are all fully and firmly in place, Check all external cables as well including connections to other devices, power outlets, etc.

Nothing should be loose or wiggly. If there are any surge protectors, power strips, check them as well.

Use a bright flashlight to carefully inspect for signs of damage: bare conductor showing, melted insulation, cracked connectors, pinched or kinked wires, browned or blackened areas, signs of liquids, swollen components, missing or loose screws, scratches, corrosion.

Intermittent problems are difficult to find and diagnose. No harm in thoroughly checking everything again. Something could turn up.