Frame rate drops to nothing

Issue Description (Required):

What’s strange is that, at first, I had no issues at all from the update. Game ran smooth as butter, just like it had for a while.

Then, out of the blue, upon load in on the Morningstar, in under a minute, I’d be hitting 1 fps. I never had any issues before this… and it’s 1 fps, then a few frames fine, then 1 fps. It’s completely unplayable.

Also, I often get an error code 3013 from the servers.

Attempted Solutions (Optional):

This is after verifying the integrity of files, reinstall, updating graphics drivers, and I don’t have any mods.

Platform (Required):

PC - Steam

[PC] PC Specifications (Optional):

Ryzen 2600, 16 GB DDR4, Radeon RX 480

Reproduction Rate:

100%

[PC] Upload Console Log & darktide_launcher.log (Optional):

console-2024-05-02-03.45.27-bee47857-085e-4297-9fb5-8933b6b2e3ac.log (149 KB)
console-2024-05-01-23.57.54-42584d6b-c0d7-451a-9bb8-a3c5f5fd1da7.log (189 KB)
console-2024-04-30-22.57.16-8e641d25-c61e-489b-8b30-2d6f9e59b140.log (98.5 KB)

[d3d12 pipeline state] creating pipeline state for <x>

my gut teels me that this could be the issue. I need to see why this happens. I will ask how i get the launcher log so you can upload this since this is part of the shader cache generation.

can you upload this file?

%AppData%\Roaming\Fatshark\Darktide\darktide_launcher.log

I think we just uploaded a new build on the experimental branch.

you can follow the instructions here

please upload a log from that that if the session contained stalls. you find them in
%appdata%/fatshark/darktide/console_logs

I should be able to see the stalls in the log then. But we have identified that there is some issue with the pso’s which is likely one culprit, usually these tends to drop off after a while since you will populate a offline cache while playing.

darktide_launcher.log (502.5 KB)

This is the before file:

When I tried to run with the beta, it crashed on launch with this log:

GUID: a7ee5b15-936f-4c45-817e-f548401d6dab
Log File:
Info Type:


this is the after experimental build file:

darktide_launcher.log (581.0 KB)

Went back to double check I had done everything correctly, and the line " disable_dstorage = true" Had been deleted from the top of the user settings document. I tried again and the line deleted itself again upon launching the game.

I tried making it read-only, and while it didn’t delete the line this time, it still crashed on launch the same way.

Crash log:

GUID: 70d1edae-f554-4d7f-a229-7abe31afa2f2
Log File:
Info Type:


it crashes when opening your .hans file in your appdata. Could you try delete this file and the .pso files (should be 2 of them) and try again.

If it happens again i would try use AMD driver cleanup tool to make sure no old drivers are lingering.

(it crashes inside the driver when trying to create a shader from the cache file)

Tried to launch the experimental version after deleting .han and .pso files. Crashed on launch

GUID: e95d79c3-b462-4faa-a61e-9ccff8fafb87
Log File:
Info Type:


Switched to the regular build, same issue remains. uploading launcher log and console log

clean driver results in same issues

darktide_launcher.log (762.1 KB)
console-2024-05-03-23.11.07-506bc5f2-930b-4bd2-8fea-72e5737ec486.log (270.5 KB)

yes, sadly same crash. I pinged AMD about this, since I can’t see what’s crashing in the driver.

Did you use any of the tools below to make sure you cleaned out the old drivers?

AMD Cleanup Utility

Display Driver Uninstaller (DDU)

I think AMD Cleanup Utility was the tool recommended by AMD to make sure you dont have older drivers lingering.

Note that you need to install the driver again, since this likely removes all GPU drivers for your AMD card.

My colleague was recommending as another tool to remove old drivers.
Display Driver Uninstaller (DDU)

If you already did something like this previously when doing a clean install don’t think it helps doing it again. I just know it’s a bit trickier to do a clean install on AMD than Nvidia (that has that as a checkbox during the driver update)

I used the Cleanup Utility as suggested, so it should be a clean install.

Are you seeing similar crashes with AMD GPUs?

I am fairly certain that the issue cropped up directly after Hotfix #37 (1.3.2)

Issue persists. It starts out fine, then within 30 seconds of moving through the morningstar, the frames start to drop. I don’t know if that indicates a memory leak of some kind. Attaching a console log, if that helps.

console-2024-05-09-22.46.25-74376565-f342-40de-ace5-a8bedcda8bb7.log (142.1 KB)

I apologize. After noticing in Afterburner that my fan speed was zero and my temps very high on my GPU, I looked inside my case.

The mounting brackets for my GPU’s fan shroud had sheared clean off, all four of them. And this is a computer that hasn’t moved in 3 years. The fan, attached to the fallen shroud, was jammed against other components.

So, long story short, I fixed it with zip ties.

Thank you for the hard work and feedback. I am sorry to have taken so much of your time.

I will note that I still have frame drops, just nothing like the one before… these can make it hard to play a precision class.

i think this may be in part to the driver downclocking the GPU and not clocking it back up in time to properly render; just a hunch.

we have several reports about perfromance issues and found a few reasons. So we are working on it.

Glad to hear that you resolved your issue.

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.