Question GPU drops to 0% in SOME games, after 5-10 minutes of gameplay

[bailey1920]

I have scoured the internet and various forums for weeks trying to find a solution for this, and nothing has worked so far.

I am experiencing major stuttering in most games, where my GPU usage drops to 0-4%, and FPS drops to 5-10 for a few seconds. I can play the games perfectly fine for roughly 5-10 minutes before the stuttering begins, then the stuttering happens about every 30-60 seconds. During the stutters, the CPU usage does go up, but only jumps from 15-20% to 30-40%. RAM never exceeds 50%.

I use an overlay to constantly monitor cpu usage, gpu usage, and temps. My temps for both CPU and GPU never go over 60C. The games are not intensive, either (e.g. Phasmophobia, Alba's Wildlife Adventure, even one 2D game...) and it does not matter whether I have the graphics on the lowest settings or the highest. Other games (Portal 2, The Sims 4, House Flipper) run flawlessly. My GPU usage playing these games typically stays somewhere between 40%-100%, I am unsure why it changes so much.

I have noticed that when the stuttering occurs, my Hard Faults/Sec will spike off the graph, well above 100, though the overall RAM usage never exceeds 50%. I have been unable to detect the issue in any benchmarking tests, unfortunately...

I am wondering if anyone has had this issue, and has any suggestions on how I can fix it. The games essentially become unplayable after that 10 minute window once the stuttering starts.

I don't understand a lot about PCs, so I am happy to try to provide any more information that may be needed.

Here are my specs:

CPU: Intel Core i7 11700F
GPU: NVIDIA GeForce RTX 3060
16 GB of RAM
PSU: AG-750M
OS: Windows 10 Home
No overclocking, and my BIOS is up to date.

Summary of what I've tried:

• Putting games on lowest graphics, mid graphics, and highest graphics settings
• Toggling v-sync in nvidia control panel and in game
• Clean install of nvidia drivers
• Putting all power settings on high performance/maximum
• Disabling background processes
• deleting discord (fixed similar issues for some people)
• windows memory diagnostics (no issues detected)

Please let me know any suggestions or if you need any more information!!

 

[Ralston18]

PSU: something got lost - no info in your post.

Use Task Manager and Resource Monitor to observe system performance.

Use both tools but only one tool at a time.

Open the tool immediately after booting.,

Observe what the system is doing and does until the displayed parameters are stable.

Then, keeping the tool window open, do some light work (emails for example) followed by browsing, running apps, and lastly gaming.

Pay attention to what system resources are being used, to what extent (%), and what is using any given system resource,

Another tool that may help is Process Explorer ( Microsoft, free).

https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer

Lastly: make and model router (or modem/router) ? The router's logs may be capturing some relevant error or issue related to game play slowdowns.

Who has full admin rights to the network router? You will need help from that person.

 

[bailey1920]

PSU: something got lost - no info in your post.

shoot, i meant to double check my model number and forgot to add it back in.
ApexGaming AG-750 M

For the router, I’m having trouble finding the model. The make is Arris. I have tried playing offline and the issue persists, would that still be able to cause issues in that case?

 

[bailey1920]

Use Task Manager and Resource Monitor to observe system performance.

Use both tools but only one tool at a time.

Open the tool immediately after booting.,

Observe what the system is doing and does until the displayed parameters are stable.

Then, keeping the tool window open, do some light work (emails for example) followed by browsing, running apps, and lastly gaming.

Pay attention to what system resources are being used, to what extent (%), and what is using any given system resource,

Okay, doing this...
I found that CPU % hovered around 20% at boot, I have a few things that open at start up but nothing too taxing... Task Manager was using the most resources. While browsing with chrome, no issues.
Playing one of the glitchy games, everything was fine again for the first 10 minutes. I played it in windowed mode because I only have one monitor and wanted to see task manager. CPU stayed 40-50%, memory never exceeded 50% still, GPU was hovering from 30-50%... Then when the stutter happened CPU dropped to like 3%! This does not show up in my normal overlay monitoring. GPU also dropped to like 2%, as usual. This lasted for a few seconds then went back to normal.

Then I switched to resource monitor. Same thing happened a few seconds later, and I checked all the graphs... EVERYTHING plummeted except memory. And hard faults/sec spiked wayyyy off the chart (way over 100. probably like, 400). I have no clue if this could mean anything... but that's what i got for now. Thanks for helping.

Edit: forgot to mention, no CPU intensive processes popped up at the time of stutter. I also tried sorting by hard faults / sec just to see, and got no significant processes to pop up to the top.

 

[Ralston18]

Nicely done!

My first thought is that there are power related issues. PSU may be nearing its designed in EOL (End of Life) or otherwise beginning to falter and fail in some manner.

However, now take a look in Reliability History and Event Viewer. Either one or both may be capturing some error codes, warnings, or even informational events just prior to or at the time of the stutters.

Look for increasing numbers of errors and varying errors. If present that, to me anyway, makes the PSU a prime suspect.

Start with Reliability History. Much more user friendly and the time line format can be very revealing.

Especially if there is some pattern involved.

 

[geofelt]

If hard faults per second is anything north of zero, you need to find out why.

A hard fault will stop the cpu dead until the fault can be resolved.
If your page file is on a HDD it takes some 40x longer than if it was on a ssd.

If this happened on just one game, I would suspect some sort of a memory leak bug.

Ram usage does not tell the whole story.
Windows tries to keep ram available for instant use if needed.
If sufficient ram is available, windows will keep code around in case it can be reused quickly.

Have task manager open while you play a suspect game.
Look at the working set of tasks that are running at the same time.
You may get a clue as to the offender.
It could possibly be chrome web pages.

 

[bailey1920]

If hard faults per second is anything north of zero, you need to find out why.

A hard fault will stop the cpu dead until the fault can be resolved.
If your page file is on a HDD it takes some 40x longer than if it was on a ssd.

Thanks for the input and ideas. I'm trying to find out what's causing the hard faults, but even using task manager I'm unable to identify any rogue processes popping up, so it's really strange to me. I am on an SSD, too. And I typically don't run anything else while I play, not chrome, discord, anything. I only recently starting running software to monitor gpu, cpu, temps, etc, because of the issues I've been experiencing. When the issues started, I wasn't using any overlays or anything... blah.

 

[bailey1920]

My first thought is that there are power related issues. PSU may be nearing its designed in EOL (End of Life) or otherwise beginning to falter and fail in some manner.

ugh, I hope not because I am not sure how to verify this and I also just got this pc under a year ago. It seems possible though.

However, now take a look in Reliability History and Event Viewer. Either one or both may be capturing some error codes, warnings, or even informational events just prior to or at the time of the stutters.

Look for increasing numbers of errors and varying errors. If present that, to me anyway, makes the PSU a prime suspect.

Hardly any errors in reliability history, the few that are there are from softwares that I installed and the uninstalled trying to find the culprit of this problem the past few days... Unless perhaps I'm looking at the wrong thing? I think I included screenshots thru the link, let me know if that doesn't work right...

For event viewer, I could only find one event around the first stutter. There were really no other events during the gameplay period and none recurring with the subsequent stutters. It was "Information" in System, Event 158 Time-Service.

"The time provider 'VMICTimeProvider' has indicated that the current hardware and operating environment is not supported and has stopped. This behavior is expected for VMICTimeProvider on non-HyperV-guest environments. This may be the expected behavior for the current provider in the current operating environment as well." Searching this on google I mostly found people having issues with their PC clock not keeping time right. I don't have this issue, by the way.

Doesn't seem particularly relevant to me, but it was all I found and maybe it's indicative of a greater issue?

Screenshots (i hope):

https://imgur.com/a/ry0pcRj

View: https://imgur.com/a/ry0pcRj

 

[Ralston18]

Just as an immediate matter of elimination:

Try running the built in Windows Troubleshooters. The troubleshooters may find and fix something.

Then try "sfc /scannow" and "dism".

References:

https://www.lifewire.com/how-to-use-sfc-scannow-to-repair-windows-system-files-2626161

https://www.windowscentral.com/how-use-dism-command-line-utility-repair-windows-10-image

However the issue seems to hardware related and possibly due to power problems.

The TPM error is interesting.

How many of those errors appear on other days?

Take a look at the details for the Kernel-Processor-Power Event ID 55. There were at least 4 (four) on 5/13.

What about preceding dates? Is Event ID 55 becoming more frequent especially starting on the 10th and thereafter?

How old the PSU? History of heavy gaming use?

 

[bailey1920]

Just as an immediate matter of elimination:
Try running the built in Windows Troubleshooters. The troubleshooters may find and fix something.
Then try "sfc /scannow" and "dism".

No luck with the troubleshooters. "dism" found no corruption.
sfc scan detected and fixed something apparently, I couldn't make any sense of the log. I rebooted and started up a game and the issue persists...

https://imgur.com/a/4oteKwQ

View: https://imgur.com/a/4oteKwQ

The TPM error is interesting.
How many of those errors appear on other days?

It looks like it's occurred once on every boot, going back to September when I got the PC.

Take a look at the details for the Kernel-Processor-Power Event ID 55. There were at least 4 (four) on 5/13.
What about preceding dates? Is Event ID 55 becoming more frequent especially starting on the 10th and thereafter?

It looks like just a boot function, they're all just info about each processor (0-15). See "Processor 5" at the top, each one is the same just "Processor 0", "Processor 1", and so on. It's consistent each date.

https://imgur.com/a/JsBNY3j

View: https://imgur.com/a/JsBNY3j

How old the PSU? History of heavy gaming use?

Assuming it was new when I bought the PC < a year ago (prebuilt from bestbuy), its about 8 or 9 months old. I BELIEVE it was built with all new parts... but the cooling system already petered out last month and I had to replace it, so it is possible that the PSU is busted too. I honestly only really used this PC to play The Sims 4 for a few hours a week up until probably March when I started getting steam games. TS4 still doesn't really give me issues, and the other games that do give me issues have given me problems since I installed them.
All that to say... I have no idea if it's been a gradual decay or not /:
Maybe an undetectable power issue? is there any other way to detect that or would I just have to try getting a new PSU?

 

[Ralston18]

If possible try borrowing a comparable PSU from a family member, friend, or co-worker.

One caveat: do not use any PSU cables that do not come with the PSU.

Plus, if you have a multi-meter and know how to use it, PSU's can be tested to some extent.

If necessary get a knowledgeable family member or friend to help.

FYI:

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Not a full test because the PSU is not under load. However any voltages out of tolerance likely indicate a faltering/failing PSU.

 

[bailey1920]

hmm, okay, got it. thank you for all the help!! i may just have to resort to taking it into a shop and seeing if they can take a look at it. i really appreciate your time and efforts!

 

[Ralston18]

You are welcome.

Do take a look at @geofelt 's comments in Post #6.

The more troubleshooting that you can try and/or do beforehand the better.

Make notes, do screen captures, look at logs, and so forth.

Just work to be a step ahead of the shop. They may try to tell you anything....

Hopefully you have "shop" choices available based on recommendations of family, friends, co-workers, etc...

 

[bailey1920]

random update, nothing interesting but figured i'd put it here just so i can have all the details in one place.

I booted up my old macbook air to see if i could at least play one of the games that doesn't work right on my PC on here. Turns out, the ONLY games in my steam library that aren't available on Mac are the exact same games that don't work right on my PC.... All the games that run fine on my PC are available for Mac lol.

I am sure it's just a coincidence, but maybe this info could be useful later.

Also, I tried a clean install of windows on my PC today, cleared all my files, the whole 9-yards, only thing i redownloaded was steam and one game, and still did not work.