Weird 2070 Super issue (Caused by Possible Microsoft Game Input Service Conflict)

[Nielo TM]

So I've been using the 2070 Super (MSI Ventus) for nearly 4 years & it worked fine till few months ago.

While playing older & newer titles, system freezes at random & the GPU fans ramps up. Upon checking Event Viewer, I always see Nvidia or video related causes.

After running through the usual drills (DDU, Win SFC, memtest86, CPU stress tests etc), the problem remained. But weirdly, running GPU stress didn't trigger any issues.

I used this excuse to get myself a 7800XT, which arrived today & its going trough it paces now, but last night, I ran even more tests on the 2070 & it passed all of em. Heaven, Valley, Superposition, OCCT GPU & VRAM test...



So what gives? Why does it only crash in actual games?

UPDATE: Check post 6 for the update

 

[NattyKathy]

So I've been using the 2070 Super (MSI Ventus) for nearly 4 years & it worked fine till few months ago.

While playing older & newer titles, system freezes at random & the GPU fans ramps up. Upon checking Event Viewer, I always see Nvidia or video related causes.

After running through the usual drills (DDU, Win SFC, memtest86, CPU stress tests etc), the problem remained. But weirdly, running GPU stress didn't trigger any issues.

I used this excuse to get myself a 7800XT, which arrived today & its going trough it paces now, but last night, I ran even more tests on the 2070 & it passed all of em. Heaven, Valley, Superposition, OCCT GPU & VRAM test...



So what gives? Why does it only crash in actual games?

How are temps on the 2070S in-game after running for awhile vs temps during stress test? Could be dried-up paste or cooler issue causing overheating during long sustained loads that stress tests are too brief to pick up on.

 

[Furious_Styles]

How long did you loop those tests for?

 

[Nielo TM]

How are temps on the 2070S in-game after running for awhile vs temps during stress test? Could be dried-up paste or cooler issue causing overheating during long sustained loads that stress tests are too brief to pick up on.

Identical. Around 75-78C. Dried thermal paste did cross my mind, but that would heat up the GPU & throttle its performance

How long did you loop those tests for?

OCCT tests were an hour each. So were valley & heaven. But ran Superposition for several hours

Also, the older game I was referring to came out 10 years. Hardly tax the system.

 

[GoodBoy]

Run GPU-Z and check the 12v inputs on the sensors page, while playing a game. Check for the minimum seen voltage.

 

[Nielo TM]

UPDATE: So the cause may not have been the GPU, but rather by Microsoft Game Input Service Conflict. See attachment. I JUST uninstalled the older version. Unfortunately the 2070 Super is no longer in my PC so can't confirm this. But at least the errors in event viewer have stopped. Keep an eye out since apparently some games may re-install the older version.

On the flip side, the new 7800XT I installed 4 days didn't crash at all. The old 2070S, Forza Horizon would have crashed within 30 min. 7800XT ran for 12 hours.

So maybe my issues were caused by faulty card AND gameinput conflicts, or just one. Or maybe the game input conflict only affects NVIDIA cards. Either way, I hope this helps someone.

 

[arestavo]

UPDATE: So the cause may not have been the GPU, but rather by Microsoft Game Input Service Conflict. See attachment. I JUST uninstalled the older version. Unfortunately the 2070 Super is no longer in my PC so can't confirm this. But at least the errors in event viewer have stopped. Keep an eye out since apparently some games may re-install the older version.

On the flip side, the new 7800XT I installed 4 days didn't crash at all. The old 2070S, Forza Horizon would have crashed within 30 min. 7800XT ran for 12 hours.

So maybe my issues were caused by faulty card AND gameinput conflicts, or just one. Or maybe the game input conflict only affects NVIDIA cards. Either way, I hope this helps someone.

I've actually recently had that exact event viewer report, with a ton of restarts of the game input service (on my 7800X3D / 4090 rig). I was thinking that they were causing some rare game freezes with Star Citizen. I think I solved it yesterday with a Nvidia driver reinstall and an AMD chipset driver reinstall.

It wouldn't shock me if it turned out to be a recent windows 11 update that borked it, but I don't have that level of software deep-diving experience.

 

[Nielo TM]

I've actually recently had that exact event viewer report, with a ton of restarts of the game input service (on my 7800X3D / 4090 rig). I was thinking that they were causing some rare game freezes with Star Citizen. I think I solved it yesterday with a Nvidia driver reinstall and an AMD chipset driver reinstall.

It wouldn't shock me if it turned out to be a recent windows 11 update that borked it, but I don't have that level of software deep-diving experience.

I recommend uninstalling the old version completely. Go to add/remove programs, look for Gameinput Service and uninstal it.

Recommend following this. There's a TY guide as well.

https://www.reddit.com/r/WindowsHel...input_service_stops_working_after_windows_11/

 

[arestavo]

I recommend uninstalling the old version completely. Go to add/remove programs, look for Gameinput Service and uninstal it.

Recommend following this. There's a TY guide as well.

https://www.reddit.com/r/WindowsHel...input_service_stops_working_after_windows_11/

I started having horrible frame pacing issues at the same time as seeing those event viewer messages (probably from a physical TPM failure). I wasn't able to resolve it using my normal troubleshooting methods (clean installs of AMD chipset drivers, Nvidia drivers, sfc /scannow, DISM online, registry cleaning, chkdsk, etc.). I fixed it all with a windows refresh. Stupid windows messed itself up on such a deep level that only a reinstall fixed it.

Either way, it is now resolved.