Someone must be able to help...?

Check Windows Update to see if an update has failed to install. Had 5 work PC's (on Win10) grind to a halt recently and when I checked it was a failed update that was trying to install over and over causing the CPU to spike.
 
The weird thing is the 30mins delay. I think the vram is potentially one to investigate as it would take certain games to stress it and the cache would fill over time. Not sure how you would test that, and again that isn't a surefire cause as you are able to run another game simultaneously without problems.
 
Just a theory but
Both games with this issue are hard on CPU & both run DX11

Not checked your other games but I believe the others are dire tx 10??
 
Good Morning folks, good news looks like the issue is solved.

I ran a 3dMark stress test last night and 90 mins and everything was fine so I was pretty confident the 980ti is working fine, so started to look into why the stutter resets after changing the resolution. Basically after fiddling around with the setting I noticed that the stutter starts when running MGS5 in 'boarderless window' (default setting), yet when I change to fullscreen there has been no stuttering yet!

So who knows what the conflict or issue is, but I am happy that it appears to be rectified somewhat.

Thanks to all of you for the help and ideas for solutions!

Cheers
 
Well that was not the end unfortunately, I just had a lucky period where there was no stutter. It came back, and I noticed that the graphics card was crashing regularly when I exited games. I did not notice this before as though the momentary black screen was normal until I saw that after this MSI afterburner was not showing a graphics! The only way to get it back was a reboot or disable / enable the GPU in device manager.

So after much more testing and retesting I think I am narrowing these problems down to the change if voltage / core clock that occurs. Nearly every time when exiting a game the card appears to crash. This ties in with being able to stress test and run 3D mark for ages without fault.

Now the problem appears to eliminate when the max power mode is enabled in the nvidia control panel. This is great, but not for my idle temps!

So questions.

What may be causing this voltage issue? Hardware problem? PSU?

It there a way to 'force constant voltage' but let the core speed change. I know MSI after burner has this option but it does not appear to work. Just looking for a way to keep the idle temps down and let the fan knock off.
 
Sounds like a stability issue if it's crashing when it changes clocks. Maybe add a touch of v core to the gpu?
 
Hi guys,

I could really do with some thoughts here after my latest testing into this stutter problem.

So I changed the PSU, this did not help.

Today I changed the monitor from a 2560 x 1440 to a 1080p. The problem is now appears to be gone. I do not have a spare 1440p to test.

So what do you guys think, a GPU problem, or monitor? I have an RMA pickup scheduled for tomorrow for the GPU but thinking of cancelling and getting a new 1440 monitor.

I have no idea how monitors work and no clue if this could be a purely monitor problem?
 
I have never heard of a monitor causing stutter that was limited to certain applications. I could understand it if everything started to stutter after a time.

My guess is it is still a fault with the GPU, possibly in the memory or memory controller. Running at 1080p uses a lot less VRAM than running at 1440p, so the new monitor may just mean the software is not using the components that are actually faulty.
 
I have never heard of a monitor causing stutter that was limited to certain applications. I could understand it if everything started to stutter after a time.

My guess is it is still a fault with the GPU, possibly in the memory or memory controller. Running at 1080p uses a lot less VRAM than running at 1440p, so the new monitor may just mean the software is not using the components that are actually faulty.

One would be able to prove this by downscaling downsampling from 1440p to 1080p on the 1080p monitor.
 
So you ran a game at 1440p in to the 1080p monitor fine? Suggests that the GPU is handling 1440p doesn't it? In any case, RMAing the GPU is probably a safe bet here.

:edit: to clarify, when I said downscaling before I meant downsampling!
 
Back
Top Bottom