Question PC crashes/restarts whenever I open a game, but on subsequent tries it works fine.

[Player_Slot_Available]

About three weeks ago, I started playing Wasteland 3, and about one week into playing it, I started getting extreme lag issues in the game (A memory leak I believe) with the game pushing my GPU to 100%. Youtube video showing the lag, I managed to fix the lag in Wasteland 3 by turning on "Hardware Accelerated GPU Scheduling" in the Windows Graphics settings. I turn it off when not playing Wasteland 3.

Now, whenever I start Wasteland 3, Monster Hunter World, Warframe, or another game the screen will either freeze after a minute or so, or the screen will just lose signal to the PC. If the former, I have to go and hold down the power button to turn it off. If the latter; the PC will restart by itself after 30 or so seconds.

After it has restarted, and I open up the game it crashed in; I will not experience any new crashes and it will work fine.
GPU does not exceed 70 degrees celcius, CPU sits at 65 degrees.
I uninstalled the current GPU drivers (529-49) with DDU and installed the previous version (527.56). There was no improvement.

This is what the event viewer says.

- System

  - Provider

   [ Name]  Microsoft-Windows-Kernel-Power
   [ Guid]  {331c3b3a-2005-44c2-ac5e-77220c37d6b4}

   EventID 41

   Version 8

   Level 1

   Task 63

   Opcode 0

   Keywords 0x8000400000000002

  - TimeCreated

   [ SystemTime]  2023-02-27T14:34:01.9177767Z

   EventRecordID 151984

   Correlation

  - Execution

   [ ProcessID]  4
   [ ThreadID]  8

   Channel System

   Computer DESKTOP-IQQE1HJ

  - Security

   [ UserID]  S-1-5-18


- EventData

  BugcheckCode 0
  BugcheckParameter1 0x0
  BugcheckParameter2 0x0
  BugcheckParameter3 0x0
  BugcheckParameter4 0x0
  SleepInProgress 0
  PowerButtonTimestamp 0
  BootAppStatus 0
  Checkpoint 0
  ConnectedStandbyInProgress false
  SystemSleepTransitionsToOn 0
  CsEntryScenarioInstanceId 0
  BugcheckInfoFromEFI false
  CheckpointStatus 0
  CsEntryScenarioInstanceIdV2 0
  LongPowerButtonPressDetected false

My specs are
GPU:RTX 3080ti (not overclocked, but I tried limiting power use to 80% in MSI afterburner to see if that helped).
CPU: Ryzen 7 5800X
MB: ROG STRIX B550-I Gaming
RAM: DDR4 Dual channel 32GB 3600Mhz CL 16
PSU: 850w Gold

 

[dudio]

That error is a power fault typically caused by the psu. What make and model is your '850 watt' power supply?

 

[Player_Slot_Available]

That error is a power fault typically caused by the psu. What make and model is your '850 watt' power supply?

It is a Cooler Master V850 SFX Gold

I installed it in December because my previous 750w PSU was insufficient for my GPU.

 

[John Chesterfield]

That's not a great unit unfortunately, it may have issues with the transient power spikes that the higher end 30 series cards tend to have, possible to swap out for a known good unit of decent quality, maybe borrow one to test?

Otherwise, have you tried running each memory stick on its own?

 

[Player_Slot_Available]

That's not a great unit unfortunately, it may have issues with the transient power spikes that the higher end 30 series cards tend to have, possible to swap out for a known good unit of decent quality, maybe borrow one to test?

Otherwise, have you tried running each memory stick on its own?

Yeah, I thought it was an issue with the RAM before I bought the new PSU. I tested the memory sticks (even had it and the CPU tested at a PC repair shop). There was nothing wrong with them.
So what would you suggest I do with the PSU? I can not afford a new PSU right now, can I power limit the GPU in MSI afterburner at 60% maybe? until I get a better PSU? And what kind of PSU would you say is sufficient for my build?

Unfortunately building your own systems is quite rare here, so it is hard for me to get ahold of parts to test with. I have to order most things online.

 

[dudio]

Do a Google search for 'psu tier list' to find ranked psu's from best to worst. Toms Hardware has a list as well, but I believe it's becoming fairly dated

 

[Player_Slot_Available]

Do a Google search for 'psu tier list' to find ranked psu's from best to worst. Toms Hardware has a list as well, but I believe it's becoming fairly dated

I have a Sffpc, so it is gonna be hard to find a good one, that is strong enough and fits in my small case. :x
(I got a NZXT H1, the non-flammable version).

 

[John Chesterfield]

Yeah, I thought it was an issue with the RAM before I bought the new PSU. I tested the memory sticks (even had it and the CPU tested at a PC repair shop). There was nothing wrong with them.
So what would you suggest I do with the PSU? I can not afford a new PSU right now, can I power limit the GPU in MSI afterburner at 60% maybe? until I get a better PSU? And what kind of PSU would you say is sufficient for my build?

Unfortunately building your own systems is quite rare here, so it is hard for me to get ahold of parts to test with. I have to order most things online.

You could try power limiting, but this is a sticking plaster at best. It's unfortunate my friend, but the only thing you can really do here is to try and get another unit, even if just a borrow or you get one from a place with a good returns policy and try it before returning it. Amazon would do for that. And art least then you'd know for sure because at present, it's not 100% it's the power supply, could also be memory too so borrowing some of that of a different brand is something else you could try.

 

[Player_Slot_Available]

You could try power limiting, but this is a sticking plaster at best. It's unfortunate my friend, but the only thing you can really do here is to try and get another unit, even if just a borrow or you get one from a place with a good returns policy and try it before returning it. Amazon would do for that. And art least then you'd know for sure because at present, it's not 100% it's the power supply, could also be memory too so borrowing some of that of a different brand is something else you could try.

I turned the power down to 60 in MSI
It did not prevent the problem from occurring. Though I crashed 6 minutes into playing Monster Hunter World, rather than only 1 minute into the game.

 

[John Chesterfield]

I turned the power down to 60 in MSI
It did not prevent the problem from occurring. Though I crashed 6 minutes into playing Monster Hunter World, rather than only 1 minute into the game.

That does suggest power is the issue here. If you do go ahead and get a different unit, run the model by us here before doing so.

 

[Player_Slot_Available]

That does suggest power is the issue here. If you do go ahead and get a different unit, run the model by us here before doing so.

Alright, I will do so. Not sure where to even start looking though, since I need something that is small enough to fit in my case, and I had a hard time last time trying to find the dimensions of the PSUs.

 

[bakory]

That error in the event viewer is a generic one, it doesnt mean anything, its just telling you: hey so your pc was on, then on again? i dont see any entry of you turning off the pc, this mean the last shutdown wasnt a regular one!

Do you have any problem if you disable gpu scheduling?

How old is your bios? have you updated it?

Before you start testing, take it apart and reseat everything.

Find and download testmem5 , TM5 + anta777 in google and you will find it, is a memory test used to check ram stability.

Then after you check your rams:
Download and install hwinfo, Heaven Benchmark and furmark to stress test your gpu, keep an eye on temperature and voltage in hwinfo while you run those tests.

 

[John Chesterfield]

The above is also good advice to check, I'm wondering about power given your reduction in power in software and the less frequent crashes but crashes nonetheless. But nothing wrong with reseating components and checking BIOS version too.

If you do get a power supply from a place you can return it easily, you don't need to mount it in the case. To just test, you can connect it all up with the new unit itself outside the case, so no danger of scratching it with screws and the like.

Edit: And ONLY use the cables that come with the new power supply. Do not use existing ones, the pinouts aren't necessarily the same.

 

[Player_Slot_Available]

That error in the event viewer is a generic one, it doesnt mean anything, its just telling you: hey so your pc was on, then on again? i dont see any entry of you turning off the pc, this mean the last shutdown wasnt a regular one!

Do you have any problem if you disable gpu scheduling?

How old is your bios? have you updated it?

Before you start testing, take it apart and reseat everything.

Find and download testmem5 , TM5 + anta777 in google and you will find it, is a memory test used to check ram stability.

Then after you check your rams:
Download and install hwinfo, Heaven Benchmark and furmark to stress test your gpu, keep an eye on temperature and voltage in hwinfo while you run those tests.

I have the same problem if I disable GPU scheduling.

I have not updated the bios since I built the PC in 2021. CPU-Z says
Brand: American Megatrends Inc.
Version: 1801
Date: 01/12/2021

Should I update it? On the Rog Strix website it says the latest non-beta version is Version 2803 from 2022/05/23.

Here is the result of TestMem5.

I am sick and don't have the energy to take it apart and reseat everything. I know you told me to do that first, but I can't do it right now, so I just did what I was able to.
I will try to run the stresstest tomorrow, or reseat everything if I am able to. I have however kept an eye on temperatures every time it crashed using MSI Afterburner. GPU is usually around 70 when it crashes, and CPU was at around 60. I will try with hwinfo next for more detailed results.

 

[johnbl]

About three weeks ago, I started playing Wasteland 3, and about one week into playing it, I started getting extreme lag issues in the game (A memory leak I believe) with the game pushing my GPU to 100%. Youtube video showing the lag, I managed to fix the lag in Wasteland 3 by turning on "Hardware Accelerated GPU Scheduling" in the Windows Graphics settings. I turn it off when not playing Wasteland 3.

Now, whenever I start Wasteland 3, Monster Hunter World, Warframe, or another game the screen will either freeze after a minute or so, or the screen will just lose signal to the PC. If the former, I have to go and hold down the power button to turn it off. If the latter; the PC will restart by itself after 30 or so seconds.

After it has restarted, and I open up the game it crashed in; I will not experience any new crashes and it will work fine.
GPU does not exceed 70 degrees celcius, CPU sits at 65 degrees.
I uninstalled the current GPU drivers (529-49) with DDU and installed the previous version (527.56). There was no improvement.

This is what the event viewer says.

- System

  - Provider

   [ Name]  Microsoft-Windows-Kernel-Power
   [ Guid]  {331c3b3a-2005-44c2-ac5e-77220c37d6b4}

   EventID 41

   Version 8

   Level 1

   Task 63

   Opcode 0

   Keywords 0x8000400000000002

  - TimeCreated

   [ SystemTime]  2023-02-27T14:34:01.9177767Z

   EventRecordID 151984

   Correlation

  - Execution

   [ ProcessID]  4
   [ ThreadID]  8

   Channel System

   Computer DESKTOP-IQQE1HJ

  - Security

   [ UserID]  S-1-5-18


- EventData

  BugcheckCode 0
  BugcheckParameter1 0x0
  BugcheckParameter2 0x0
  BugcheckParameter3 0x0
  BugcheckParameter4 0x0
  SleepInProgress 0
  PowerButtonTimestamp 0
  BootAppStatus 0
  Checkpoint 0
  ConnectedStandbyInProgress false
  SystemSleepTransitionsToOn 0
  CsEntryScenarioInstanceId 0
  BugcheckInfoFromEFI false
  CheckpointStatus 0
  CsEntryScenarioInstanceIdV2 0
  LongPowerButtonPressDetected false

My specs are
GPU:RTX 3080ti (not overclocked, but I tried limiting power use to 80% in MSI afterburner to see if that helped).
CPU: Ryzen 7 5800X
MB: ROG STRIX B550-I Gaming
RAM: DDR4 Dual channel 32GB 3600Mhz CL 16
PSU: 850w Gold

EventID 41
when the machine reboots with out completing a shutdown then event id 41 will be recorded during the next boot sequence.

in general you would ignore this log entry and focus on why your machine rebooted without a clean shutdown.
IE power failure, bugcheck, motherboard reset of cpu.

most memory problems would result in a bugcheck being recorded and a minidump file being created. generally you would see a error code for an access violation in these cases.

Issues that are due to power to a GPU can cause a motherboard protection circuit to reset the CPU and no minidump would be created. (first step is to remove any overclocking and retest)

 

[bakory]

I have the same problem if I disable GPU scheduling.

I have not updated the bios since I built the PC in 2021. CPU-Z says
Brand: American Megatrends Inc.
Version: 1801
Date: 01/12/2021

Should I update it? On the Rog Strix website it says the latest non-beta version is Version 2803 from 2022/05/23.

Here is the result of TestMem5.

I am sick and don't have the energy to take it apart and reseat everything. I know you told me to do that first, but I can't do it right now, so I just did what I was able to.
I will try to run the stresstest tomorrow, or reseat everything if I am able to. I have however kept an eye on temperatures every time it crashed using MSI Afterburner. GPU is usually around 70 when it crashes, and CPU was at around 60. I will try with hwinfo next for more detailed results.

Yea you should update your bios, you are missing a lot of AGESA updates

 

[Player_Slot_Available]

Yea you should update your bios, you are missing a lot of AGESA updates

I have now updated the BIOS.
I have also reseated everything. The CPU temperature is better after reseating.

I tried running MHW as usual afterwards, and I usually get 60FPS stable, but now it ran at 30 from the start, and just went lower and lower until this happened.

After I clicked on OK, the PC shut itself off.
It is hard to see but in the top right corner it says
CPU Utilization: 29%
GPU Utilization: 11%
GPU temperature: 37c

I had not undervolted or limited the power to the GPU when running this test.

 

[bakory]

I have now updated the BIOS.
I have also reseated everything. The CPU temperature is better after reseating.

I tried running MHW as usual afterwards, and I usually get 60FPS stable, but now it ran at 30 from the start, and just went lower and lower until this happened.

After I clicked on OK, the PC shut itself off.
It is hard to see but in the top right corner it says
CPU Utilization: 29%
GPU Utilization: 11%
GPU temperature: 37c

I had not undervolted or limited the power to the GPU when running this test.

Check event viewer in system for the crash log.

Saw a reddit post saying that he fixed this crash by disabling windows HDR, also dont forget to install and update your chipset drivers.

 

[Player_Slot_Available]

Check event viewer in system for the crash log.

Saw a reddit post saying that he fixed this crash by disabling windows HDR, also dont forget to install and update your chipset drivers.

Unfortunately, I am now unable to turn the PC on.
The CPU cooling does not seem to work. The temperature swiftly rises when I turn on the PC and by the time I am able to log in, it has reached 110 degrees celcius. Then it crashes. One of the cooling tubes is cold the other is warm so I assume it works, the fan spinning, and the metal plate is touching the cpu. I do not know what is wrong now. But it seems I made it worse by reseating everything.
I attempted to reseat everything once more but it did not help.

Other than begging the only PC shop in the region to take a look at my PC and break the rules (they only accept PCs bought from them) I don't know what I can do. This is beyond me.

 

[bakory]

Unfortunately, I am now unable to turn the PC on.
The CPU cooling does not seem to work. The temperature swiftly rises when I turn on the PC and by the time I am able to log in, it has reached 110 degrees celcius. Then it crashes. One of the cooling tubes is cold the other is warm so I assume it works, the fan spinning, and the metal plate is touching the cpu. I do not know what is wrong now. But it seems I made it worse by reseating everything.
I attempted to reseat everything once more but it did not help.

Other than begging the only PC shop in the region to take a look at my PC and break the rules (they only accept PCs bought from them) I don't know what I can do. This is beyond me.

What fan cooler do you have? is your cpu is reaching 100c before loading windows then something is very wrong with the fan cooler like not good contact between cpu and heatsink, fan cooler connected in the wrong spot.

 

[Player_Slot_Available]

What fan cooler do you have? is your cpu is reaching 100c before loading windows then something is very wrong with the fan cooler like not good contact between cpu and heatsink, fan cooler connected in the wrong spot.

It is an AIO cooler. The one that came with the NZXT H1 case.
I suspected it might not be in contact properly so I unscrewed the heatsink and then held it pressed up against the CPU to make sure it was in contact. Just in case there was something wrong with the mounting. The temperatures hovered at around 85c while in bios when I held it like that.
Last week it used to idle at around 30-40c.

Edit: I let it cool for a day, and tried turning it on again. I recorded it.
When I got to the bios it climbed to 80. When I rebooted the PC and logged in it just crashed, I assume because it climbed up above 100 before it managed to get to desktop.
Youtube recording