• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

TheRealCrusaderKing

Corporal
46 Badges
Oct 20, 2019
33
3
  • Crusader Kings II
  • Stellaris - Path to Destruction bundle
  • Victoria 2: A House Divided
  • Hearts of Iron III
  • Darkest Hour
  • March of the Eagles
  • Europa Universalis III
  • Hearts of Iron IV: Death or Dishonor
  • Crusader Kings III
  • Stellaris: Lithoids
  • Stellaris: Ancient Relics
  • Prison Architect
  • Hearts of Iron IV: Expansion Pass
  • Imperator: Rome
  • Cities: Skylines Industries
  • Stellaris: Distant Stars
  • Cities: Skylines - Parklife
  • Stellaris: Apocalypse
  • Stellaris: Humanoids Species Pack
  • Hearts of Iron IV: Expansion Pass
  • Cities: Skylines - Green Cities
  • Cities: Skylines - Mass Transit
  • Hearts of Iron IV: Together for Victory
  • Stellaris: Leviathans Story Pack
  • Stellaris: Digital Anniversary Edition
  • Hearts of Iron IV: Cadet
  • Stellaris
  • Cities: Skylines - After Dark
  • Mount & Blade: With Fire and Sword
  • Cities: Skylines
  • Europa Universalis IV
  • Victoria 2
Hi,

Game keeps crashing the display driver (screen cuts to black, system still operational in the background, needs power cycle to fix).

I am running: R7 3700X, RX 5700 XT, Radeon driver version 19.9.3.

A quick Google shows that this issue has been around at least six months, most common with AMD users but some reporting it on Nvidia cards as well. So far updating drivers has not helped any of them.

Any acknowledgement of the issue? ETA on a fix?

Really enjoying the game!

Cheers.

Edited to add more detail on system config.
 
Last edited:
Quick update: installed the latest Radeon driver (19.10.2), issue still occurring.

Seems to be at random - sometimes I can play for hours without it happening, sometimes half an hour.

I love the game but it's frustrating to have to power cycle my machine periodically to continue playing.

Anyone at Paradox willing to admit that this is a known issue?
 
A quick Google shows that this issue has been around at least six months, most common with AMD users but some reporting it on Nvidia cards as well.
Specifically relating to this game? Or PDS games? Or just a general issue? Because while I'm not necessarily on top of everything all the time, I don't recall other such reports with this game.


DXDIAG is a program you run from a command prompt or the Windows start menu 'run' dialog box (or 'search programs' in Windows 7 or later). After running it will open a window and start collecting info with a progress bar in the lower-left corner. When it completes click the 'save all information' button and save it to a file then attach that file here.

Please attach your Documents/Paradox/Imperator/pdx_settings.txt.
From your Documents/Paradox/Imperator/logs/ folder, attach system.log .

Please report the temperatures of your CPU and video (GPU)
1/ before starting the game up and
2/ after say 30 minutes' play.
If your PC's vendor has given you no way of reading temperatures, install
SPEEDFAN http://www.almico.com/sfdownload.php
 
Specifically relating to this game? Or PDS games? Or just a general issue? Because while I'm not necessarily on top of everything all the time, I don't recall other such reports with this game.

Specifically with this game, no other games PDS or otherwise. Reported to Reddit in April:

https://www.reddit.com/r/techsuppor...play_driver_crash_in_imperator_rome_amdkmdap/

Two people in that thread with identical symptoms to me. Then here again, possibly the same person posting to Steam community, also gets a reply from someone with the same issue:

https://steamcommunity.com/app/859580/discussions/1/1676938384253422105/

DXDIAG is a program you run from a command prompt or the Windows start menu 'run' dialog box (or 'search programs' in Windows 7 or later). After running it will open a window and start collecting info with a progress bar in the lower-left corner. When it completes click the 'save all information' button and save it to a file then attach that file here.

Please attach your Documents/Paradox/Imperator/pdx_settings.txt.
From your Documents/Paradox/Imperator/logs/ folder, attach system.log .

Attached as requested :).

Please report the temperatures of your CPU and video (GPU)
1/ before starting the game up and
2/ after say 30 minutes' play.
If your PC's vendor has given you no way of reading temperatures, install
SPEEDFAN http://www.almico.com/sfdownload.php

I use Ryzen Master for the CPU and GPU-Z for the GPU. I've stress tested both and built the machine with thermals in mind because of where I live. Under heavy load for 60 minutes the CPU does not exceed 78C and the GPU does not exceed 90C. The CPU idles at 30 - 35C and the GPU only a couple of degrees above that. Ambient is in excess of 25C.

Symptoms aren't right for that anyway - the game continues playing and I can alt+tab out and back into a Youtube window and hear everything. The display driver is crashing. I've even tried updating chipset drivers in case it's an issue with PCIe. Nothing seems to work.

It happened again today, I can't find a rhyme or reason to it. I just got done playing for six hours without any issues but before starting that session I'd been playing for maybe 20 minutes before the driver crash. It's really strange.

Appreciate your help!

Cheers.
 

Attachments

  • TheRealCrusaderKing-DxDiag.txt
    80,7 KB · Views: 49
  • pdx_settings.txt
    4 KB · Views: 10
  • system.log
    696 bytes · Views: 9
Last edited:
We only offer support here in the forums and our ZenDesk tickets system. We don't remotely have the time to monitor Reddit, Twitter, Facebook, Discord, Steam, and the numerous other channels around.

Under heavy load for 60 minutes the CPU does not exceed 78C and the GPU does not exceed 90C
So there is your problem, 80c is the redline. I thought I'd see that Vsync was off, but to get 90c with it on is something! I see your desktop refreshRate is a modest 60hz, what fps are you seeing in the game? You might try using a utility to cap it to 60 also; while I'm not familiar with Catalyst I guess it might have such a function? Or use fraps.

Also Clean out all air vents it has (compressed air cans are good for this). Make sure airflow to them is not blocked by anything like walls, books or whatever.

If that doesn't drop the temps enough, you may have a faulty fan or a heatsink/thermal paste problem.
 
We only offer support here in the forums and our ZenDesk tickets system. We don't remotely have the time to monitor Reddit, Twitter, Facebook, Discord, Steam, and the numerous other channels around.

Fair enough. I (wrongly) assumed the information must have made it to Paradox.

So there is your problem, 80c is the redline. I thought I'd see that Vsync was off, but to get 90c with it on is something! I see your desktop refreshRate is a modest 60hz, what fps are you seeing in the game? You might try using a utility to cap it to 60 also; while I'm not familiar with Catalyst I guess it might have such a function? Or use fraps.

The 90C I reported was the junction temperature under a stress test. It doesn't come close to that in game. For instance when I ran Ark: Survival Evolved for a couple of hours it only hit 82 and that's by far more demanding than any PDX title (used to crash my old system by overheating the card). The junction is hotter than what is usually reported as "GPU" temp so I take that as the worst case.

Also Clean out all air vents it has (compressed air cans are good for this). Make sure airflow to them is not blocked by anything like walls, books or whatever.

I do this regularly (with compressed air), all vents are clear. There are 2x140mm 1800rpm fans pushing 128CFM as intakes and 1x120mm 2100rpm pushing 80CFM out the back, this is also the fan I put on the heatsink. I know what I'm doing :p.

If that doesn't drop the temps enough, you may have a faulty fan or a heatsink/thermal paste problem.

Unlikely, the heatsink is rated to double the TDP of the CPU and I'm using Arctic Silver for the paste. The GPU has 3 fans, all working fine, and a massive heatsink.

It's not overheating. When the card overheats the system doesn't just keep running like that, it shuts down.
 
For clarity's sake, after posting the above I fired up GPU-Z and then ran Imperator: Rome for 30 minutes. I loaded a save where I had a dozen stacks and set them to patrol borders so they'd walk around the whole time.

Before:

before.png


After:

after.png


CPU taken same time as the after screenshot:

CPU.png


Ambient is 27C.

I used stress test and Ark results because they're much worse than a PDX game will ever be and I thought that would demonstrate the point but I guess that it's better to just get you the real data.
 
The 90C I reported was the junction temperature under a stress test. It doesn't come close to that in game.
I see, sorry I didn't read carefully enough.

You definitely do not get this with any of our other gams that i see under your name here?

Do these crashes leave entries in the Windows Event viewer?
 
You definitely do not get this with any of our other gams that i see under your name here?

Admittedly haven't tried them all (on this system), but I definitely don't with HoI4.

It's never happened in any scenario but I:R, and then sporadically at random. The screen doesn't need to be particularly 'busy', it's not happening during autosaves, I can't find a pattern at all.

Do these crashes leave entries in the Windows Event viewer?

Nope, not in Application or System logs anyway. I can see when I'm doing the power cycles but no events related to display or I:R in the leadup. My assumption is that the display driver is crashing because the only symptom is the monitor thinking it's been disconnected and going black, then to sleep. Everything else continues as normal.
 
Last edited:
Just had it happen again immediately on starting a new campaign. Managed to alt+tab before the monitor went to sleep and checked error.log under Documents\Paradox Interactive\Imperator\logs:

[17:47:12][gfx_dx11_swapchain.cpp:195]: D3D error (D:\jenkins\workspace\sulla-steam\clausewitz\pdx_gfx2\DX11\gfx_dx11_swapchain.cpp:195): hr returned: The GPU device instance has been suspended. Use GetDeviceRemovedReason to determine the appropriate action.

[17:47:23][gfx_dx11_base_buffer.cpp:60]: D3D error (D:\jenkins\workspace\sulla-steam\clausewitz\pdx_gfx2\DX11\gfx_dx11_base_buffer.cpp:60): hr returned: The GPU device instance has been suspended. Use GetDeviceRemovedReason to determine the appropriate action.

[17:47:23][gfx_dx11_master_context.cpp:426]: Failed creating constant buffer

Hope this helps!
 
Same issue, same setup, only on Imperator: Rome. When the 5700 XT first launched, this was happening to me with a whole heap of other games, the drivers were super unstable and even now they might need some more ironing out. If Paradox can help with this that'd be great, but perhaps it's something AMD needs to correct on their end.

Edit: I'll upload my info too if it's of any use. Also can confirm; temps are not the issue.

And as far as entries for IR in Event Viewer go, I found these:
Code:
Event 1002

The program imperator.exe version 1.0.0.0 stopped interacting with Windows and was closed. To see if more information about the problem is available, check the problem history in the Security and Maintenance control panel.
 Process ID: 3e18
 Start Time: 01d58fa97ca5ed44
 Termination Time: 4
 Application Path: G:\Games\Steam\steamapps\common\ImperatorRome\binaries\imperator.exe
 Report Id: b7d32ca4-76f1-4eaf-be93-d092a1e33136
 Faulting package full name:
 Faulting package-relative application ID:
 Hang type: Unknown
Code:
Event 4101, Display

Display driver amdkmdap stopped responding and has successfully recovered.
 

Attachments

  • Colesy_DxDiag.txt
    95,8 KB · Views: 48
  • crashes.rar
    3,3 MB · Views: 10
Last edited:
Okay, at this point I'm ready to say this is a card or driver issue; why just with this game now when it used to affect others I cannot say for sure, but it's not unusual given that different games use different functions in the driver/card combo in different ways.

By all means raise this with our QA folks in Bug Reports, hopefully they have an RX 5700 AMD video setup to confirm this on.

Good luck!
 
Thanks Andrew,

I've posted there and included error logs from my and Colesy's posts.
I'd honestly be surprised if you got a meaningful response at all. From what I understand, AMD doesn't really address major issues specific to games unless there's a very large outcry about it. Example being Destiny 2 when Zen 2 launched.
 
Same issue on Radeon 380 and only on this game, (no other Paradox game or others that are so more taxing than this, I see same temperature values for other games) with the latest drivers. It can go for decades but eventually it crashes the GPU driver and progress is lost since last save, so annoying.
 
I'd honestly be surprised if you got a meaningful response at all. From what I understand, AMD doesn't really address major issues specific to games unless there's a very large outcry about it. Example being Destiny 2 when Zen 2 launched.

I posted it in Bug Reports here :).

Same issue on Radeon 380 and only on this game, (no other Paradox game or others that are so more taxing than this, I see same temperature values for other games) with the latest drivers. It can go for decades but eventually it crashes the GPU driver and progress is lost since last save, so annoying.

Thanks for confirming the same thing on an older card - I've added your comment to the bug report.
 
I posted it in Bug Reports here :).
Oof. My brain omitted the part when Andrew said "with our QA folks". Either way, it might be an issue worth asking AMD about, though I'd be skeptical given what I mentioned in my previous post.
 
I agree with your previous post; I think Paradox are much more likely to care about the issue than AMD.

Also now that we have a report from someone using an earlier generation card, also only in I:R, I think Paradox have the onus tbh.
 
Problem still occurs with the latest Radeon driver (19.11.1)