(Nigh) Instant Crashing on character select screen

Warhammer 40,000: Darktide Technical Support
1

Issue Type (Required):

Crash - During Play

Issue Description (Required):

Greetings! Recently I got my girlfriend (a step towards a gaming PC) PC upgrade, featuring a shiny new RX 6600 8gb ddr6 GPU, that will be horribly bottlenecked by the i3 3770 (4 Cores, 8 Threads 3.4GHz) I know, I know, we will upgrade, eventually, but for now, the game ran surprisingly well!
First week of playing> no notable issues, other than some light stutters when larger hordes spawn (to be expected) and I assumed fixable. Now a slightly bitter issue was crashing upon opening and trying to claim hestiaā€™s blessings, so i set out to troubleshoot the stutters and the crashes with her.
We tried these things, in (approximately!) this order:
changing some AMD base settings, like antilag, framegen, radeon sync, etc. No notable difference.
Changing worker threads, (lowering to 3 from 5) and then going to 6, here we got another hestia crash so clearly that wasnt it.
We were monitoring the CPU and GPU usage while in game and loading, and (surprisingly) the CPU wasnt even at 100% usage all the time, the amps and temp was all good, literally nothing out of the ordinary.
We then looked up the crash error we were getting, and it is the classic ā€œDXGI ERRORā€ which we looked up and learned is a GPU bound error, so thats strange!
We then tried doing some reddit fixes (i know dumb idea) and disabled CFG for darktide, (which we then undid when the really bad crashing started happening) and at the same time deleteing the darktide folder, then verifying files, as it is shown in the ā€œhow to resolve gpu crashesā€ forum article made by julia.
Now, it might be important to say, that we did NOT disable steam cloud saves before doing this, because we forgotā€¦
We also fiddled with some advanced radeon options, like surface format optimization and tessellation mode, however, i cant for the life of me remember wether that was in the same launch as when this crashing started or one before. Either way, radeon resets any settings we change after each crashā€¦
And this is where the crashing on character screen comes into play.
It is STILL the same error message. The game loads the intro cinematics and logos, we skip using space, get to the part where the character is rendered and doing the idle animation, and in 10 seconds (every time, 10 seconds) or if we try to press the ā€œstartā€ button, the game crashes, with the same error message as before, except it now does that EVERY time.
There is a way to postpone this, if we press enter REALLY fast after skipping the cutscenes, we can get into the escape and settings menu, which allows us to fiddle with settings, but no help, the game just autocrashes no matter what settings we do.

I tried to be as detailed as possible, because this is a really complex chain of events we ā€œtroubleshootedā€ (holy moly that went wrong) and im trying to give as many helpful bits as possible, because to me, the whole thing does not make sense.

Me and my girlfriend called it a night by uninstalling the radeon driver (we tried launching, now it wont even load the body of the character as the game crashes, great.)
We will try installing it tomorrow, and if nothing works we will try a clean reinstall of darktide, however, the whole situation does not add up to make sense for me.
First of all, CPU usage and GPU usage is low, and the whole thing is happening on character select.
What is even weirder is that its the SAME error message we have been trying to fix before too, but it persists, even after deleteing the entire Darktide folder from %appdata% as mentioned before!
Maybe something went wrong with verifying the game files?
Thanks, hope we can resolve it because I would love to keep playing darktide with my GF, these couple of weeks were great, and also if there are any recommendations to fix general stutters too that would also be appreciated (like, lowering or increasing working threads, some settings recommendations.)
Cheers!

[PC] Do You Use Mods? (Optional):

No, I donā€™t use mods

Reproduction Rate (Required):

Constant (100%)

Upload Supporting Evidence (Optional):

darktidecrash22.PNG
darktidecrash22.PNG404Ɨ331 112 KB

darktidecrash22.PNG
darktidecrash22.PNG404Ɨ331 112 KB

2

Are you overclocking your RX 6600? dxgi_error_device_removed generally means the OS lost contact with the GPU, typically because the GPU crashed. If youā€™re OCing, try removing it, or if youā€™re not, see if you can under clock it a little bit temporarily for testing.

Beyond that, you may want to try and reinstall the gpu drivers. Windows is sorta dumb, so to make sure itā€™s less dumb download the drivers, download DDU. disconnect from the internet, run DDU to remove all the old drivers, reboot, stay disconnected from the internet, install the drivers you downloaded, and then reconnect to the internet.

1 Like
3

Sorry to hear this @MaxxSoreson

Could you please copy and paste a crash report (weā€™re after the GUID string) so we can look up the logs and take a closer look?

Good suggestions above, please try disabling any overclocks (if any), and using a driver removal tool such as DDU and reinstalling fresh drivers may help!

You also mentioned deleting your AppData and verifying file integrity - however performing a clean reinstallation could be worth a try, which should let us rule out an issue with the game installation itself:
https://support.fatshark.se/hc/en-us/articles/360034118394--PC-How-to-Perform-a-Clean-Re-Installation-of-Vermintide-2-or-Darktide

1 Like
4

Thank you for your fast response Lev, really cool!

We got the GUID for you, also tried reinstalling the drivers (we uninstalled them just by control panel yesterday, tried reinstalling, and the game still crashes, we will try DDU after this)

GUID: fbe93e49-c793-43c4-be75-7013f915da42
Log File:
Info Type:

[Engine Error]: HRESULT: 0x887a0005 = DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

At D:\a\d\rel_engine_darktide\release\2025_02_19_Patch1.7.x\runtime\d3d12_render_device\d3d12_debug.cpp:1548 in function stingray::aftermath::d3d_assert

5

In fact, here is another crash report, we got plenty to spare.
We tried going into settings and turning on / off portrait rendering (thats also something we messed around with to try to fix stutters) But every time the game crashes it seems that settings are reset because everything at the bottom of interface is always ON

GUID: 88ffae30-1dbf-4ffe-b2f8-766af7e063ff
Log File:
Info Type:

[Engine Error]: HRESULT: 0x887a0005 = DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

At D:\a\d\rel_engine_darktide\release\2025_02_19_Patch1.7.x\runtime\d3d12_render_device\d3d12_debug.cpp:1548 in function stingray::aftermath::d3d_assert

6

Thanks @MaxxSoreson

I shared the crash reports with the development team, and they noted that the logs show many failed pso compilations. They also recommended to use a dedicated driver removal tool such as DDU and to try a fresh GPU driver installation afterwards.

Please let me know how it goes!

7

Well, we did a DDU, proper, without internet connection while the driver is being reinstalled, and it just crashed againā€¦ We selected a full wipe on pretty much everything AMD radeon related. I am honestly starting to get pretty worried hereā€¦ She is currently reinstalling darktide, its a pretty slow connection, it will likely only be downloaded later at night.
Any other ideas for when it is reinstalled? Try running an older driver?
Thing is, we already tried completely uninstalling the drivers and running the game like that, and the same crash (except worse) occured. I just have no idea why this is happening, the game worked just fine before all of this messā€¦

In case you need it here is another crash codeā€¦

GUID: 11c9b550-b53d-4527-9b51-70b4fe987a1c
Log File:
Info Type:

[Engine Error]: HRESULT: 0x887a0005 = DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

At D:\a\d\rel_engine_darktide\release\2025_02_19_Patch1.7.x\runtime\d3d12_render_device\d3d12_debug.cpp:1548 in function stingray::aftermath::d3d_assert

8

Ahh hm okay. One of our developers suggested to try this next, to see if anything changes:

  1. Press Windows Key + R
  2. Enter %appdata% and select ā€˜OKā€™
  3. Navigate to the Darktide directory e.g. \AppData\Roaming\Fatshark\Darktide
  4. Delete the following 3 files:
  • shader_cache.hans
  • shader_library.pso_lib
  • state_stream_library.pso_lib
  1. Launch Darktide as normal

Please let me know :slight_smile:

9

Alright, we will try. Though I am pretty sure this was the exact directory we tried deleteing and then verifying the steam files before. So this time, delete specific things, and dont verify. Gotcha.
I will let you know how this goes after the game finished reinstalling and it still doesent work (i bet 700 aquillas it wont)
Thank you for the incredible and fast support so far, one of the few companies I can name that take such issues seriously! Ill keep you updated.

2 Likes
10

Alright so here we go, the reinstall itself didnt do anything, BUT what you recommended, deleting these specific files did make the game construct some new shaders, and WE DID NOT CRASH ON CHARACTER SELECT!
Now, we did crash pretty much 2 minutes later when fiddling with settings, since motion blur was reset, we tried turning that back off and yeah the game crashed immediately as the button was pressed.

Now, since we can seemingly get in game, what is the next step in making the game playable?
Meaning eliminating stutters and making sure these crashes dont happen. At least not this often.
Before when we were playing together she only crashed when she claimed a penance, and stuttered in the lobby in general.
So, knowing the PCā€™s specs (btw its 16 gigs of ddr4 ram im pretty sure, except if its ancient-ass ddr3 ram)
what are some settings we can SAFELY mess around with, in radeon and in game to help with these crashes and stutters?

I wanna make it clear, the job isnt done yet, we can load in past character select now, but its still crashing more than it did before. Clearly some setting has been reset, and I want to avoid any more chaos, so what are some recommended settings?
Yes, we are using FSR 3.1 in game, and currently everything is turned off in the radeon driver.

Thank you for getting us this far!

This might be a different error code so i will post it just to be sure>

GUID: 87d9fb75-8c81-48b2-890c-21ca49cd344a
Log File:
Info Type:

[Engine Error]: HRESULT: 0x887a0005 = DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

At D:\a\d\rel_engine_darktide\release\2025_02_19_Patch1.7.x\runtime\d3d12_render_device\d3d12_debug.cpp:1548 in function stingray::aftermath::d3d_assert

11

Sorry to hear the crashes are persisting! :frowning: hopefully heading in the right direction though.

I spoke with the development team again, and they recommended to try running a RAM test at this point, to make sure there isnā€™t a hardware fault.

You can check your RAM for errors using the Windows Memory Diagnostic by:

  1. Press the Windows key + R
  2. Enter mdsched.exe and select ā€˜OKā€™
  3. When ready, select the ā€˜Restart now and check for problemsā€™ option
  4. Allow the scan to complete

If errors are produced, we recommend that you seek advice from a hardware specialist.

12

Hello, so, we ran the tests and checked the results, no issues have been detected in the RAM.
Did the dev team say that this could be a ram issue? I am not sure what this test is for, physical ram or Vram? Want to know because im terrified something is wrong with the gpu (i picked it, yikes)
Also we messed around with putting a larger paging file size, this was my idea to improve performance.
Anyway, we surely are heading in the right direction because the game now doesent autocrash after 10 seconds of seeing a model, so the shader cache was buggy in some way, but we still need some info on what settings and tweaks we can try to resolve the crashes.
Surely these crashes are not caused by the same issue, since we crashed 100% of the time on character select, now it only crashes more often than it did before all of this mess.
What would you recommend we should set the CPU worker threads at for example, with such a bottlenecked GPU?
What about ingame settings that could help? Etc

13

If you havenā€™t yet, in the graphics options near the bottom, set the rag doll number to 3, and make sure the rag doll interaction is set to off.

14

Okay,looong update time, this is legit extremely weird, today we spent pretty much half the day troubleshooting, and we drew some interesting conclusions, but sadly its a one step forward one step back sitch.

So we tried launching today, and the game crashed on character select again. So. One step back.
We literally didnt change anything since last time, so this just happened on its own, seemingly?

Anyway, went on a troubleshooting spree, using every forum and also chatgpt (was actually quite useful)

-We noticed the power draw spikes from 20 to almost 100 before the crash.
-We edited in a line in regedit that allowed the time before the crash to be extended, to a 30s wait time (instead of the 0 it was before) this gave us a totally new error message:

GUID: 4358c53f-6da9-4d82-acc5-4ebf16017cda
Log File:
Info Type:

[Engine Error]: Deadlock detected while waiting on renderer thread. Update was not called for 16.1 seconds. You need to supply a crash dump or a crashify link in the report since we need both logs and crashdump to investigate this.

[Crash Link]:

crashify://4358c53f-6da9-4d82-acc5-4ebf16017cda

Quite interesting, have no idea what it means. ChatGPT was also a bit confused but ultimately pointed at either GPU driver failure or pipeline fault (we knew this much before i think)

Now, we fiddled with all sorts of things, because as far as i can recall, not 100% sure but, i think the 100% crash on character screen started happening when we disabled/enabled the CFG for darktide. (I checked, mine is overwritten to be off. I, weirdly dont remember doing this, please ask a dev if this is default or wtf is going on)
So we tried enabling, disabling CFG, deleting the 3 recommended files again, verifying steam files and going back to a radeon driver that is 23.11.1 (2023 release, properly, with deep DDU)
None of this helped, but i noticed that the 3 deleted files are actually not there, its only
shader_cache.hans that is present, the other 2 were gone.
SO i tried sending her copies of my 3 files to replace, we did, but it didnt help.

We did many combinations of many things, undervolting the GPU in manual tuning (we first lowered it by 15%, nothing, when we lowered it completely, it resulted in crashing at the intro cinematic when your wonderful shark mascot crashes into the screen, how poetic.)

We then stumbled across files in %localappdata%\steam\htmlcache\ShaderCache , led to localappdata via robot.
At this point out of curiosity, we opened the shadercache and noticed 4 data files, and 1 index file, and saw that one data file, data_1 was last modified today.
We then, out of pure desperation, hunted down all of these steam shader cache files that were modified today, or since the crashes happened.
We also turned off steam overlay at the same time, if this is relevantā€¦

And guess what, we could get in gameā€¦ for 2 minutes again. She crashed after buying a duelling sword and while clicking the upgrade button.

Launch again, this time we just idled, literally no movement, aboard the morningstar, we crashed around 1 minute of time in game.

Since i was using chatgpt as a smart notepad and telling it everything inbetween happenings, it actually gave some hopeful words, saying that we are now battling a fragile pipeline / instability, and not constant hardware failure.

So really to me there are 3 key points / questions

  • Is this related to CFG or is it just a coincidence / we did something else too and the combination of things messed us up
  • Are the crashes shader cache related
  • Are they power draw related
  • Or both?

I hope this new error message we started getting after increasing the time in the registry editor to 30s can provide devs / support with a proper lead, we are kinda starting to see patterns, and clearly steam shader files might have changed stuff, but it could have also been the steam overlay we disabled?
I feel so confused but also like we are making progress at the same time, and equally terrified this could be a PSU issueā€¦ pleaseā€¦ help? I really want to play darktide with my girlfriend, I bought it as an anniversary giftā€¦

15

Instead of undervolting, have you tried reducing the power limit? Undervolting too far can also cause instability, but reducing the power limit will help lower the draw on the power supply.

Speaking of power supply, what is it rated at? If itā€™s too small it could lead to DXGI_ERROR_DEVICE_REMOVED/DXGI_ERROR_DEVICE_HUNG.

16

So, we tried limiting AND giving it more power, from power saver / eco mode / undervolting / power limit in AMD, darktide peaks at 60w and still crashes.
We tried ultra performance and increased power limit, and the game goes over 100 ish and crashes all the same.

The PSU is 500w, and not a noname one eitherā€¦ cant remember the name but it should be solid, combined with the old and less power hungry CPU, it should cap out at around 300/350 max, but its not even reaching that much and crashing.

Also gotta note that the game crashes even if she just AFKs in the mourningstar, so heavy load / power issue might not be it? Im also being hopeful here.

17

500 watt should be plenty for a 6600 and an i7 3770. Iā€™m still thinking itā€™s hardware and not software, but I could be wrong.

Is the Bios up to date? Only other thing coming to mind is the 6600 is a PCI-e 4.0 8X, and the mobo is PCI-e 2.0 16X and maybe thereā€™s an issue with the negotiation. Can you try forcing the mobo to pci-e 2.0, or 1.0 instead of auto negotiating?

Do there games play just fine? Can you run furmark for 10 minutes without it crashing? How about some 3dmark benchmarks.

Can you try swapping parts with your system to rule some things out?

Unfortunately itā€™s a 12 year old mobo/CPU/memory and thereā€™s a possibility it has developed instability.

18

Alright so, we tested it with battlefield 1 (as far as i understand it uses a much different shader/rendering technique, so totally different pipeline) and sadly its a similar sitch, it crashes after a few minutes of gameplay, similarly to darktide and sometimes even on loading maps and in the second of loading in.
The crash report is the following.

bf1crashreport
bf1crashreport1920Ɨ1080 187 KB

We will try to run some 3d graphic testing on it.
Interestingly, we found that DDU is detecting old drivers, that were hidden and unable to be cleared by it (we ran ddu multiple times now, only now did i pay attention to the logs:)

ddudrivers
ddudrivers1920Ɨ3413 732 KB

We went into device manager, there were no other drivers displayed, until we checked to view hidden drivers.

hiddendrivers
hiddendrivers1920Ɨ3413 1.07 MB

There was no option to disable, so we just deleted everything here. Still getting the error message mentioned above.
The only thing that has helped so far is clearing out those shader caches, both the ones from the appdata darktide directory and the local appdata steam ones, so i really dont know.
The BIOS, i dont know how up to date it is, i will have to check in with my uncle who did the installation of the new gpu/ssd (i talked to him today, he has no better clue than anyone here)

I will try the following things:
furmark bench, examine aida64 more closely to see if its a PSU issue (I still think that is really unlikely, it would be the motherboard first i think)
We can try downloading a separate windows to the HDD and running darktide on a clean windows and new everything just to rule out system problems but this is difficult, and close to a last resort thing, because the logistics of me doing this alone are complicated, i am currently not living with my GF and would have to learn how to install a clean windows, and no my uncle cant helpā€¦
And i will wipe everything, literally every shader cache on the PC that is related to steam or games and let them rebuildā€¦ if anyone has a suggestion on what to target lmk.

@FatsharkLev Please, is there any kind of lead, from the error messages i sent, the deadlock one, or anything the dev team relays that could help? We are getting really desperate hereā€¦

1 Like
19

With the battlefield 1 results being the same, itā€™s sounding like some sort of hardware issue to me. Could still be driver, but I doubt it at this point.

As much as it sucks, it may be time for the mobo/cpu/ram upgrade. If you have a microcenter near you they have some decent bundles.

20

You think its a motherboard issue and not GPU? I do have to say we got the GPU second hand, though it was not used for mining, it was used in a gaming rig in an internet cafe for a couple of years until it got upgraded. What makes you think its a mobo/cpu/ram issue?
Also, note that battlefield is saying the same thing in the error message> something wrong with the DRIVER, which can be rooted back to the problemā€¦
Just, how is that logical? The game ran completely fine for over 2 weeks! It did have some crashes but only when talking to hestia, and not always. This guaranteed crashing happened after we fiddled with the settings as mentioned aboveā€¦ it just wouldnt make sense for it to be like ā€œalright i will physically dying because they turned this setting onā€ i dont knowā€¦