Crashes/Freezes/Shutdowns - how can I find out the issue?

+
Greetings,



Its weird, I have a new PC since 3 month ( 4090 + i9 14900k ) and I played cyberpunk for like 40 hours without any issues.

But starting yesterday the shit hits the fan.. and I don't know if its a cyberpunk issue or a hardware issue because it doesn't only happen on Cyber.

All Drivers are up2date, also, I haven't changed anything since it was still running smoothly



There are 3 scenarios that randomly happen whenever I load into my game ( no mods installed, also not overclocked hardware )



Scenario 1:

Once the loading screen is done and I press spacebar to enter the game it just closes itself with the "send report" popup window

The report shows me randomly one of these 3 errors:


Error reason: Assert Expression: <Unknown> Message: Watchdog timeout! (120 seconds) File: E:\R6.Release\dev\src\common\engine\src\engineWatchdog.cpp(198)


or


Expression: EXCEPTION_ACCESS_VIOLATION (0xC0000005)

Message: The thread attempted to read inaccessible data at 0x48.


or


Message: Gpu Crash for unknown reasons! Callstack here is probably irrelevant. Check if Breadcrumbs or Aftermath logged anything useful.

File: E:\R6.Release\dev\src\common\gpuApi\src\dx12\gpuApiDX12Error.cpp(42)"



Scenario 2:


This scenario happens 80% of the time

After loading screen I press spacebar to enter the game, I see the city and ready to go but the game is already frozen, sound still continues ( sometimes sound gets scuffed )

can't do anything now, opening task manager or closing game doesn't work anymore, same with pressing alt+ctrl+del for 10+ seconds to reboot, I have to manual shutdown the computer with power button



Scenario 3:



This only happened once so far, after loading into the city the PC just shutdown itself and rebooted

I also tried Unigine Heaven benchmark for testing, the computer freezed at scene 12. I tried Unigine Superposition Benchmark, it crashed before it even starts

Rocket league runs okay but two times the game also closed itself, which never happened before

The only game that runs without any problems is Genshin Impact, but I don't know if that requires much hardware power



How can I find out what the issue is.. is there any good benchmark or tool to find the issue?


I'm at a loss as to what else I could try
 
i can try to help a bit:

first of all, Whats your PSU?

1. Watchdog is a monitor tool for intel, generaly its an unstable overclock that causes it. It can be without an OC nowdays tho since most motherboards has settings that boost the performance of a cpu and with intels own boosts it can become unstable.

Expression: EXCEPTION_ACCESS_VIOLATION (0xC0000005) is generaly a memory issue but it can be something else too (yea i know, its so stupid with error codes that can pretty much be anything). Do you have XMP enabled? what ram speed are you using?

This only happened once so far, after loading into the city the PC just shutdown itself and rebooted

I also tried Unigine Heaven benchmark for testing, the computer freezed at scene 12. I tried Unigine Superposition Benchmark, it crashed before it even starts
Also points towards a unstable cpu. I finally got my 13900k to stop random shutdowns by lowering my ram speeds from 7200 too 7000 and disable real time memory tuning (the motherboard try too make the ram go faster pretty much) My memory controller in my cpu is kinda bad so it just couldent handle it. its kinda hard to give specifics here since i cant se it myself but you could try to use bios default settings to se if it helps, will probably be abit slower but if it works it points towards some hardware thats just not stable. Also the gpu could be factory overclocked and instable. You using any overlays to check temps and so on?
 
i can try to help a bit:

first of all, Whats your PSU?

1. Watchdog is a monitor tool for intel, generaly its an unstable overclock that causes it. It can be without an OC nowdays tho since most motherboards has settings that boost the performance of a cpu and with intels own boosts it can become unstable.

my PSU is a 1000W be quiet! Straight Power 11

Expression: EXCEPTION_ACCESS_VIOLATION (0xC0000005) is generaly a memory issue but it can be something else too (yea i know, its so stupid with error codes that can pretty much be anything). Do you have XMP enabled? what ram speed are you using?

i don't know what XMP means, and RAM i have 6000MHz Kingston Fury Beast, not sure what the ram speed is, guess i have to check that in BIOS?
I just did a AIDA64 benchmark for RAM and Cache, its just a test version but here are the results:
Screenshot 2024-02-04 185849.png

But jesus I got scared, when the test was running it made some really strange squeaking noises and the fans were spinning like crazy, i could not abort the test because it used so much performance that i couldn't do anything else like switching to browser



Also points towards a unstable cpu. I finally got my 13900k to stop random shutdowns by lowering my ram speeds from 7200 too 7000 and disable real time memory tuning (the motherboard try too make the ram go faster pretty much) My memory controller in my cpu is kinda bad so it just couldent handle it. its kinda hard to give specifics here since i cant se it myself but you could try to use bios default settings to se if it helps, will probably be abit slower but if it works it points towards some hardware thats just not stable. Also the gpu could be factory overclocked and instable. You using any overlays to check temps and so on?

i have some tools for temps like hwmonitor and aida but the temps are weird, when i do CPU stresstest it jumps from 30°C to 90°C within seconds, and when i stop the test its back at 30°C also within seconds, is this even possible that it cools down so fast?
I think 90°C is high but its still fine, googling my CPU it says it can have up to 100°C while gaming on ultra
 
sadly i never changed keybindings, i was used to use CTRL for dashing.. i still tried to set it back to default but it did not help
 
Your earlier screenshot shows that your RAM was at 4800 MHz. If it's advertised at 6000 MHz XMP, then XMP isn't on.

This is like tech support challenge. Too bad it's the middle of the night here. It sounds like hardware configuration (bios/drivers), assembly (how it's put together) or health prob.

I can look at the pieces you posted over the week, but I'm not a tech support pro. If you have a buddy/hiree who builds and fixes PCs for a living to come over and have a look, they might be a better option.
 
Your earlier screenshot shows that your RAM was at 4800 MHz. If it's advertised at 6000 MHz XMP, then XMP isn't on.

This is like tech support challenge. Too bad it's the middle of the night here. It sounds like hardware configuration (bios/drivers), assembly (how it's put together) or health prob.

I can look at the pieces you posted over the week, but I'm not a tech support pro. If you have a buddy/hiree who builds and fixes PCs for a living to come over and have a look, they might be a better option.

Yea well i don't have such a buddy, I am living in the middle of nowhere :rolleyes:

Also i checked my Bios, and yes XMP was turned off, but DDR5 Auto Boost was enabled, also there was a option with "high bandwidth" and "low latency" i also turned that off, don't even know what that does.

Also this Gigabyte Z790 board has "Gigabyte Perfdrive" options in the BIOS, by default "optimization" is enabled, other choices would be unleash, instant 6ghz, Spec Enhance and last one would be E-core Disable

Unleash sounds like a heavy overclock i will not touch that, same with 6ghz, don't really need more cpu power
Perhaps i try Spec enhance and E-core disable tomorrow to test if the issue is gone, but i doubt it
 
I set cyberpunk to low details and I also set the maximum FPS at 60, because if I keep it unlimited it would just generate 500+ FPS in low details and the components would be at their maximum again

And I actually was able to play for 15 minutes. Then i set graphics higher and higher and then it freezed again, had to press power button to force a shutdown.

So it only happens when the PC is working hard. Could it be the power supply? But 1000W should be enough for that system

But i also tried to create a new savegame on ultra details, and the freeze also happened there while being in the character creation, thats weird to me because i don't think character creation needs much power

But at least we can now narrow down the cause somewhat
 
Man i tried everything and nothing works..

Atleast the bluescreens i managed to fix by removing nvidia drivers with that DDU tool in safe mode and installing it new.

But instead of a death bluescreen it still freezes but the game closes and drops me back to desktop.

And somehow it got worse now, i can't play for more than 2 minutes until it crashes.

im getting gpuApi\src\dx12\gpuApiDX12Error.cpp(42) error and the windows debug thingy is saying something about "nvlddmkm" which is nvidia file but i have installed the newest driver after removing it with DDU, can't be a driver issue

I also runned a OCCT stability stress test for CPU, Linpack, Memory, 3D Standard and Power, 1 hour each, no errors and no crashes at all

I am about to give up and just uninstall the game.. its sad but i have no fucking clue how to solve this and thats freaking me out
 
Man i tried everything and nothing works..

Atleast the bluescreens i managed to fix by removing nvidia drivers with that DDU tool in safe mode and installing it new.

But instead of a death bluescreen it still freezes but the game closes and drops me back to desktop.

And somehow it got worse now, i can't play for more than 2 minutes until it crashes.

im getting gpuApi\src\dx12\gpuApiDX12Error.cpp(42) error and the windows debug thingy is saying something about "nvlddmkm" which is nvidia file but i have installed the newest driver after removing it with DDU, can't be a driver issue

I also runned a OCCT stability stress test for CPU, Linpack, Memory, 3D Standard and Power, 1 hour each, no errors and no crashes at all

I am about to give up and just uninstall the game.. its sad but i have no fucking clue how to solve this and thats freaking me out
Annoying :(
Did you at least, contact the support?
> "contact us" button
 
Yea i did wrote the support but no response yet. Also i think its not a Cyberpunk issue, i just played a other game ( dead by daylight ) with unreal engine and the game crashed after 5 minutes.

Weird thing it also crashed my discord and the nvidia control panel is gone from the quickbar at the bottom right windows taskbar.
Cyberpunk also kills my discord app but not the nvidia panel, man what in the world is this?

The game error was this:
Fatal error: [File:Unknown] [Line: 202] Unreal Engine is exiting due to D3D device being lost. D3D device was not available to assertain DXGI cause.
 
Yea i did wrote the support but no response yet. Also i think its not a Cyberpunk issue, i just played a other game ( dead by daylight ) with unreal engine and the game crashed after 5 minutes.

Weird thing it also crashed my discord and the nvidia control panel is gone from the quickbar at the bottom right windows taskbar.
Cyberpunk also kills my discord app but not the nvidia panel, man what in the world is this?

The game error was this:
Fatal error: [File:Unknown] [Line: 202] Unreal Engine is exiting due to D3D device being lost. D3D device was not available to assertain DXGI cause.
From what I read on the net, "Unreal Engine is exiting due to D3D device being lost" error seem to be almost like "EXCEPTION_ACCESS_VIOLATION (0xC0000005)" error on Red Engine, be caused to all kind of things, be caused to all sort of things, from faulty hardware to antivirus, drivers or overclock/underclock... :(
 
Yea i did wrote the support but no response yet. Also i think its not a Cyberpunk issue, i just played a other game ( dead by daylight ) with unreal engine and the game crashed after 5 minutes.

Weird thing it also crashed my discord and the nvidia control panel is gone from the quickbar at the bottom right windows taskbar.
Cyberpunk also kills my discord app but not the nvidia panel, man what in the world is this?

The game error was this:
Fatal error: [File:Unknown] [Line: 202] Unreal Engine is exiting due to D3D device being lost. D3D device was not available to assertain DXGI cause.

The fact your issues aren't solely limited to CP2077 is a clear indicator that the issue is on your end.

This D3D error points to your GPU. D3D stands for Direct3D. Going over your original post, I would also assume your GPU is the issue here. You've already stress tested everything and only your GPU test causes your PC to crash (I admit I skimmed through the thread and may have missed stuff). That's not to say there aren't other possible hardware issues at play here but I would definitely look at your GPU first.

When you crashed here, were you able to just start a game right away? Did you have to reboot your PC? What happened after the crash? I'm trying to determine whether your GPU is crashing or completely "disconnecting".

Also, you don't strike me as the person who assembled this PC. Am I correct? Is it pre-built? If so, can you indicate the exact model?
 
The fact your issues aren't solely limited to CP2077 is a clear indicator that the issue is on your end.

This D3D error points to your GPU. D3D stands for Direct3D. Going over your original post, I would also assume your GPU is the issue here. You've already stress tested everything and only your GPU test causes your PC to crash (I admit I skimmed through the thread and may have missed stuff). That's not to say there aren't other possible hardware issues at play here but I would definitely look at your GPU first.

When you crashed here, were you able to just start a game right away? Did you have to reboot your PC? What happened after the crash? I'm trying to determine whether your GPU is crashing or completely "disconnecting".

Also, you don't strike me as the person who assembled this PC. Am I correct? Is it pre-built? If so, can you indicate the exact model?

Yes correct its a pre-built Computer with these specs:

Operating System: Windows 11 Home 64-bit
System Manufacturer: Gigabyte Technology Co., Ltd.
System Model: Z790 GAMING X AX
BIOS: F9a (type: UEFI)

Processor: Intel(R) Core(TM) i9-14900KF (32 CPUs), ~3.2GHz
Endorfy Navis F360 cooler

Card name: NVIDIA GeForce RTX 4090
Current Mode: 3440 x 1440 (32 bit) (165Hz)

32GB (2x16GB) DDR5 AMD/Intel 6000MHz Kingston Fury Beast

1000W be quiet! Straight Power 11


Yes i can start a game right away after the crash, games like rocket league or diablo 4 are running without any problems, but cyberpunk and now also dead by daylight are giving me hard time.
I also don't understand why these crashes also kill my discord app.
How do i know if the GPU completely disconnected? Because its confusing me that the nvidia control panel symbol in the bottom right disappeared after the crash

The last days i had bluescreens instead of game crashes, but i fixed that with a clean new install of the nvidia driver after deleting the old one with DDU.
Before the clean driver install it was like this:
I started the game, i walked around some seconds or in best case some minutes, then it froze, but sound was still there, and after like a minute or so it went black and the PC did a shutdown + restart.

And now i start game, walk around for some seconds/minutes, it freezes again, and after 10 seconds the screen turns black and game + discord crashes. But i could just restart it to experience the same thing again and again


Ah yea previous in the thread i mentioned that i had problems with Heaven and Superposition Benchmarks, i tried them again today, and i had no problems at all, it did not crash.
 
Yes correct its a pre-built Computer with these specs:

Operating System: Windows 11 Home 64-bit
System Manufacturer: Gigabyte Technology Co., Ltd.
System Model: Z790 GAMING X AX
BIOS: F9a (type: UEFI)

Processor: Intel(R) Core(TM) i9-14900KF (32 CPUs), ~3.2GHz
Endorfy Navis F360 cooler

Card name: NVIDIA GeForce RTX 4090
Current Mode: 3440 x 1440 (32 bit) (165Hz)

32GB (2x16GB) DDR5 AMD/Intel 6000MHz Kingston Fury Beast

1000W be quiet! Straight Power 11


Yes i can start a game right away after the crash, games like rocket league or diablo 4 are running without any problems, but cyberpunk and now also dead by daylight are giving me hard time.
I also don't understand why these crashes also kill my discord app.
How do i know if the GPU completely disconnected? Because its confusing me that the nvidia control panel symbol in the bottom right disappeared after the crash

The last days i had bluescreens instead of game crashes, but i fixed that with a clean new install of the nvidia driver after deleting the old one with DDU.
Before the clean driver install it was like this:
I started the game, i walked around some seconds or in best case some minutes, then it froze, but sound was still there, and after like a minute or so it went black and the PC did a shutdown + restart.

And now i start game, walk around for some seconds/minutes, it freezes again, and after 10 seconds the screen turns black and game + discord crashes. But i could just restart it to experience the same thing again and again


Ah yea previous in the thread i mentioned that i had problems with Heaven and Superposition Benchmarks, i tried them again today, and i had no problems at all, it did not crash.

From the errors you are getting and everything that's happening (including the Nvidia panel disappearing), it looks like your GPU is briefly stopping. This can be caused by quite a few things.

It could also explain why Discord crashes if your sound is being carried through your GPU. Your GPU crashes/disconnects briefly and now Discord is getting all kinds of errors because it's supposed to receive sound from a GPU that's just suddenly disappeared. It really depends on how things are set up on your end. If you go into Discord -> Voice & Video and you'll see the devices Discord uses to receive and output sound. If you're going through your GPU, it'll show up as "Nvidia High Definition Audio" if I recall correctly.

Soooooooo.... how comfortable are you with opening your PC? Is it a "not happening even if my life was on the line" type of situation? I don't know if you have any experience doing that.

My current thinking is that your GPU might not be seated properly or has come loose. It's also possible that the power connectors powering it have come loose or, as has happened quite a few times in the past with the 4090, burnt up. In an ideal scenario, it's just a loose card/connectors.

With that said, I don't want you to go rummaging in there if you are not comfortable/have zero experience doing so or if there is a chance this might void your warranty. If you're willing to do it, please make sure there is nothing stopping you from doing that in your warranty. It's very likely ok but better safe than sorry.

If you know how to do this and are comfortable with it, go ahead and unplug the power from your PC and have a look at that card and those connectors.
 
From the errors you are getting and everything that's happening (including the Nvidia panel disappearing), it looks like your GPU is briefly stopping. This can be caused by quite a few things.

It could also explain why Discord crashes if your sound is being carried through your GPU. Your GPU crashes/disconnects briefly and now Discord is getting all kinds of errors because it's supposed to receive sound from a GPU that's just suddenly disappeared. It really depends on how things are set up on your end. If you go into Discord -> Voice & Video and you'll see the devices Discord uses to receive and output sound. If you're going through your GPU, it'll show up as "Nvidia High Definition Audio" if I recall correctly.

Soooooooo.... how comfortable are you with opening your PC? Is it a "not happening even if my life was on the line" type of situation? I don't know if you have any experience doing that.

My current thinking is that your GPU might not be seated properly or has come loose. It's also possible that the power connectors powering it have come loose or, as has happened quite a few times in the past with the 4090, burnt up. In an ideal scenario, it's just a loose card/connectors.

With that said, I don't want you to go rummaging in there if you are not comfortable/have zero experience doing so or if there is a chance this might void your warranty. If you're willing to do it, please make sure there is nothing stopping you from doing that in your warranty. It's very likely ok but better safe than sorry.

If you know how to do this and are comfortable with it, go ahead and unplug the power from your PC and have a look at that card and those connectors.

i will give it a try, I had to plug in the graphics card myself when the computer was delivered anyway; it was delivered separately due to transport safety.

But i think i fixed the crash issue with the "dead by daylight" game, i just had to set "preferred graphic processor" to high performance nvidia processor in the "manage 3D settings" menue from the nvidia control panel, not sure if its really the fix but i am playing since a hour now without a crash. have to check if this simple thing also worked with cyberpunk
 
Your power supply is multirail - 4 rails. Can somebody explain what it actually means when you plug graphics card to power supply.
 
i will give it a try, I had to plug in the graphics card myself when the computer was delivered anyway; it was delivered separately due to transport safety.

But i think i fixed the crash issue with the "dead by daylight" game, i just had to set "preferred graphic processor" to high performance nvidia processor in the "manage 3D settings" menue from the nvidia control panel, not sure if its really the fix but i am playing since a hour now without a crash. have to check if this simple thing also worked with cyberpunk

It's possible it's just that but, frankly, it's very unlikely. it should default to that from the get go. It really shouldn't go to your on-board graphics if you have a dedicated card. it may have defaulted to your on-board because your GPU briefly disconnected.'

Alternatively, and I don't really suggest you do that but it's really your prerogative, give your PC a shake or your desk really when you're gaming. If it's a loose connector/card, it'll crash.

Again, I don't recommend this as it's purposefully causing a crash and shaking possibly loose hardware looser and all but it's a quick check.
 
Atleast the bluescreens i managed to fix by removing nvidia drivers with that DDU tool in safe mode and installing it new.
You mean this?
Windows blue screen error.png
Experienced this too a couple of times, and decided to bring the system back to the dealer to have it examined. There, it was determined both the GPU and CPU to be faulty, and were replaced with new ones.

We asked what the cause was for them to become faulty, as both were only 3 months old, but they couldn't say. Seemingly, the manufacturers never tell anyone anything. Faulty components just get replaced as long they fall under warranty.

So, if yours is brand new, despite you being able to fix the blue screens, would recommend that you simply offer up your system with your dealer to be examined, hence so the components'll be replaced with new ones.
 
Top Bottom