[Workaround found !] Random crash mid mission, not seem related to GPU or memory lack- 3/4 games

Warhammer 40,000: Darktide Technical Support
1

Continuing the discussion from KNOWN ISSUE: Crashing at Random, Due to GPU-Related Errors:

Hello all,

Issue Description:
I’m still facing 3/4 games to random crash, back at my desktop …

I’ve monitored my performances, i didn’t not see some kind of critical value reached on GPU / Ramm / CPU nor I/O disk… the code of the game seem to crash without any ressource consumption issue, or it’s to short to be monitored by the system… can’t explain that

i’m playing on lowest specs for ALL settings except for the worker tread still a tnine/ putting it to 1 doesn’t change a things for me…

Error Displayed (If Applicable):
None => back to desktop following a few second freeze, warmly welcomed by the crash report tool…

latest Crash Report :
GUID: 0c3f51a1-22e5-46af-9edd-e4a138d174bf
Log File:
Info Type:

Steps to Reproduce / Mission Name :
Totally random, got it on various game / situation. Never on the ship deck however !

Platform:
Microsoft Store

ID :
Ymir77

Approx. Time of Issue & Timezone:
Various, see logs (UTC)

Reproduction Rate:
Often (<75%)

Upload Supporting Evidence:
Nothing more than logs…

My setup :

Win 11 x64 22H2 build 22621.963
Nvidia geforce RTX 3600 (latest → pilote ver 527.56)
AMD Ryzen 5 3600 6-Core Processor 3.60 GHz
16 GB ramm well synced with the cpu
Installed on SSD 500 Go
wide monitor 3840x1080 144Hz

I can play with the same fluidity whith Mid specs / mid ray tracing, the crash do not happen faster…
i play in low spec expecting to have less issue, but in fact i do not truly see any improvement.

I tried all the Know how in GPU / Random memory crash, without real improvement.
i can only blame the poor game code at this point that seem to have some critical flaw…

thx for help
Reports after this post

2

Some console logs :

console-2022-12-19-12.34.35-e283e963-21ba-42bd-989d-63efb78ae33b.log (348.7 KB)
console-2022-12-19-13.15.31-5e32cc90-a8e3-46ee-b606-3c4950af4c81.log (186.8 KB)
console-2022-12-19-13.45.08-eb22b744-9825-44c8-a596-7543e437bd57.log (149.9 KB)
console-2022-12-19-15.41.44-0c3f51a1-22e5-46af-9edd-e4a138d174bf.log (164.5 KB)
console-2022-12-19-11.41.20-f2b7b10a-fab7-44c5-b6bd-3430ee81880a.log (194.6 KB)

3

launcher logs :

darktide_launcher.log (639.2 KB)

4

i found interesting logs :

15:56:42.528 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (60692), last frame (60691).
15:56:42.536 [Lua] INFO [AdaptiveClockHandlerClient] Filtering out frame dt of 0.059108, more than 2.000000 times old max dt of 0.029129
15:56:42.539 [Lua] INFO [PlayerUnitDataExtension] Panic resolved, took 0.011 seconds.
15:57:02.262 warning: [WwisePlugin] Wwise::post_event. Failed to trigger event `play_power_sword_hit_husk` for object id: 260, name: `unknown`
15:57:25.952 [DRED] DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG
15:57:25.957 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (63406), last frame (63377).
15:57:26.953 [DRED] <<DRED Hang>> DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG <</DRED Hang>>

15:57:26.953 [AFTERMATH] Device status: GFSDK_Aftermath_Device_Status_PageFault
15:57:26.954 [AFTERMATH] Page fault information:
15:57:26.954 [AFTERMATH] Faulting GpuVA: 0xf3bdc000
15:57:26.954 [AFTERMATH] Resource desc:
 						size: 0
 						width: 0
 						height: 0
 						depth: 0
 						mip levels: 0
 						format: DXGI_FORMAT_UNKNOWN
 						Buffer Heap: false
 						Static Texture Heap: false
 						RtvDsv Texture Heap: false
 						Placed Resource: false
 						Was destroyed: false
15:57:27.348 <<Crash>>HRESULT: 2289696773 = DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

so the game cannot handle a frame desync :
“Received frame (63406), last frame (63377)”

GPU get king of a virtual saturation => direct X panic => frame desync => game crash …

following the honnest spec i have it seem that the game have an issue with some randoms action triggering suche abnormal behavior ingame, maybe the usage of direct X ?

Since i’m not a develloper i leave here my feedback hoping to have some clue a least from the support !
Good luck on this rocky trip to stabilize the game.

Expecting patch soon :smiley:

5

BTW auspex seem to have also some issue, i see theses kind of error in many of my logs but not always related to crash

warning: [WwisePlugin] Wwise::post_event. Failed to trigger event `play_device_auspex_scanner_minigame_progress_last_husk` for object id: 139, name: `unknown`
15:54:50.511 warning: [WwisePlugin] Wwise::post_event. Failed to trigger event `play_device_auspex_scanner_minigame_progress_last_husk` for object id: 153, name: `unknown`
15:55:43.245 warning: [WwisePlugin] Wwise::post_event. Failed to trigger event `play_device_auspex_scanner_minigame_progress_last_husk` for object id: 153, name: `unknown`
6

New crash 1st game directly in the first minutes :

GUID: eb5f009f-9f73-4b8c-9c3e-30cf53ae55fd
Log File:
Info Type:

Tried to deactivate the direct X pilote loading failure detection that sometimes is too sensible by modifying the TdrLevel in registry to 1 ( Testing and debugging TDR during driver development - Windows drivers | Microsoft Learn) without improvement…

Exactly the same error :

21:12:43.327 [DRED] DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG
21:12:43.330 [Lua] INFO [PlayerUnitDataExtension] **Panic started**. **Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (23462), last frame (23433).**
21:12:44.327 [DRED] <<DRED Hang>> **DXGI_ERROR_DEVICE_REMOVED**: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG <</DRED Hang>>

21:12:44.330 [AFTERMATH] Device status: GFSDK_Aftermath_Device_Status_PageFault
21:12:44.330 [AFTERMATH] Page fault information:
21:12:44.330 [AFTERMATH] Faulting GpuVA: 0xccfec000
7

new crash 2nd try to play …

impossible to play the last games … it crash every mission !

Tried to cap the FPS to 60 in NVidia control pannel, because it was confirmed that the game wasn’t able to handle large monitor… (very good news in the next gen era…)

image

GUID: 3f3781c0-90da-4444-bf4f-42cba1bccd0d
Log File:
Info Type:

So no improvement so far…
still pursuing my journey alone trying to debug myself …

SAME ERROR Again :

22:21:34.325 [DRED] DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG
22:21:34.331 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (93580), last frame (93552).
22:21:35.325 [DRED] <<DRED Hang>> DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG <</DRED Hang>>

My several logs are all filed with : Panic started. Received state package from the future, this is likely because we haven’t run any fixed frames yet or had a major stall. Received frame (XXX), last frame (XXX).

But the on time it cannot recover = game crash !
Direct X crash
the game crash
and back to desktop…

I don’t have any clue so far, so i hope that someone from support will see this post since i see other topic like mine and with the same error on their logs !

exemples :

10:03:43.783 [Lua] INFO [AdaptiveClockHandlerClient] Filtering out frame dt of 0.520305, more than 2.000000 times old max dt of 0.037094

13:45:56.170 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (27611), last frame (27610).

10:43:45.235 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (9722), last frame (9721).

08:50:59.042 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (10733), last frame (10705).
8

@Support I hereby invoke you in the name of the emperor :smiley:

image
image1920×1080 129 KB

1 Like
9

I think i fixed my frame panic issue … playing several hours without crash (never happened before…)
I have still a lot of them in the logs however they quickly resolve each time and the game can normally process then.

Logs of today :
console-2022-12-20-08.55.33-6d47f533-1a35-45fc-8ba5-ee71a322b83b.log (672.1 KB)

Still got a lot of

08:58:04.568 [Lua] INFO [PlayerUnitDataExtension] Panic started. Received state package from the future, this is likely because we haven't run any fixed frames yet or had a major stall. Received frame (508900), last frame (508899).
08:58:04.588 [Lua] INFO [AdaptiveClockHandlerClient] Filtering out frame dt of 0.056239, more than 2.000000 times old max dt of 0.035200
08:58:04.599 [Lua] INFO [PlayerUnitDataExtension] Panic resolved, took 0.072 seconds.
  1. First I capped the FPS at 60 in the NVIDIA Panel AND in the game setting.
  2. Then i also deactivated the user setting within the file. in fact some where still in true state despite being deactivated in game setting.

Source : https://www.reddit.com/r/DarkTide/comments/z9a2fj/psa_rtx_is_always_on_despite_what_options_you/

Credit goes to Solivaguz who was the one who figured this out.
It Seems like Ray Tracing is always enabled, although it seems to be off in the ingame settings.
1.Set all graphics options to off/low in the game. (DLSS on balanced or w/e)
2.Close the game and open %appdata%\Fatshark\Darktide\user_settings.config using notepad.(C:\Users*YOUR_USER*\AppData\Roaming\Fatshark\Darktide)
3.Make sure the following settings are set like this:

rt_checkerboard_reflections = false
rt_light_quality = false 
rt_mixed_reflections = false 
rt_particle_reflections_enabled = false 
rt_reflections_enabled = false 
rt_shadow_ray_multiplier = 0 
rt_transparent_reflections_enabled = false 
rtxgi_enabled = false 
rtxgi_scale = 0

4.Once this is done you should notice a big FPS bump that’s more inline with what you’d expect with your setup. You can tweak any settings you want ingame, just don’t touch the RTX ones.

Theses both fix seem to let my game work…

But as other people said it’s not an acceptable solution.
People should not have to fine tune their configuration to allow the game to process as expected…
people do not have the knowledge nor the necessity to to date as you are the product dev owner. you have a hardware detection service to tune the games setting ! why it’s so messed up !

I took me several try and a loot of setting test to find a NOT NORMAL solution.

Hope you can learn from my case for the next patch it seem that i’m not the only one why frame panic !

If this can help other people … @FatsharkJulia

3 Likes
10

I’ve upgraded my settings, i’m now in Mid-high graphics settings without crash anymore.
The flaw was clearly in the frame regulation that the game cannot manage properly.

Logs if needed :
console-2022-12-20-12.14.59-604ec8a7-be28-4b38-8868-d247dbe90729.log (442.5 KB)

@FatsharkLev this may interess you also.
have a good day

1 Like
11

@Abriel, oh wow, nice one!

Thank you for the detailed write up here! I’ve shared this with the development team.

12

@Abriel
Just sharing some developer feedback regarding this -

For users modifying the rtxgi_scale or any of the rt settings manually to the tech priest note in the community update for patch 1.0.14. The guide they are following isn’t really a good one as it will make some levels pitch black without turning off raytracing any more than turning it off in the menu does.

The early community guide to disable Raytracing is no longer valid since the 1.0.14 update.

The capping of the framerate is probably what is helping them by either rearranging the timings so that the issue doesn’t happen or with fewer frames they likely get fewer gpu hangs.

3 Likes
13

I tried limiting the game to 60 fps with the user_ini edits but it doesnt help T_T

14

bro can you attach here your user_settings.cfg ? i’ve got the same processor and the 3070 and got the same error

15

thanks for the feedback, in fact i restored the settings back.

However, i don’t know why i need to cap the 60 fps in both GAME setting menu and NVidia Pannel control (sad to not be able to have more T_T).
Don’t know why but both need to be caped to stop crash !

Frame panic are still observed inside the logs, however it is managed and recovered, but not sure if i can’t crash anymore, but it’ stable enough to play !

thx anyway !

1 Like
16

it’s not really the setting ini that helped me since i restored when i improved my graphic quality this morning but more caping in Nvidia control pannel (see my screen before in the thread) AND in game setting through the menu in game !

anyway i share the setting if it kind provide some clue.
user_settings.config (3.1 KB)

17

Sigmar H Emperor can you STOP PROPAGATING BAD INFO with the rtxgi_scale crap and all the RT settings, these were debunked but they went viral before the devs can do damage control.

If anyone with an ounce of tech knowledge had read the previous tech stuff before this crap went viral on Reddit because of some dude on Steam tried to get his posts pinned with awards and crap, they would have noticed the talk of RT for baked lighting. It’s an obscure thing to read I admit and I wish the tech devs had debunked this crap SOONER. That Steam user you credited is mostly full of it, just regurgitating config tweaks from other users or using some obvious stuff like diving power of 2 by half.

There are legit bugs to fix and all this broken telephone nonsense just creates noise in the social echo-chamber, making it virtually impossible to share legit trouble shooting steps.

18

Can we PIN this here, Reddit, steam? This bad info is STILL being propagated and is just making things worse.

I suggest for future stuff like this a dev come out with a debunk post within 12-24 hours before it snowballs out of control.

19

Well … it’s not stupid if it’s working

20

What’s working exactly?