Multiple CUDA/RTX/Vulkan application crashing with Xid (13,109) errors

I’d like to report that the crashes with some vkd3d titles (FH5, etc.) on pre-Turing hardware, seem to be fixed in the latest beta 555.52.04 as well as 550.40.63 vulkan dev beta.

In addition, after this regression fix, it was finally possible to test the Xid error fixes for FH5 implemented in 550.40.07-550.54.14.

So, I can report that FH5 didn’t crash once in hours of gameplay. It was finally possible to use unlocked framerate with a mix of Ultra and High graphical settings (~90-100fps) on my 1080ti. :)

Driver Version: 555.52.04

I tried The last of us and the XID 109 crash is still there. It seems like the first run with new drivers the game seems to work.
The next time you start the game it will start crashing and will continue to do so until you update the drivers again.

[tis jun 11 11:04:24 2024] NVRM: GPU at PCI:0000:01:00: GPU-c527f869-4b6b-fb00-51ea-f36233874170
[tis jun 11 11:04:24 2024] NVRM: Xid (PCI:0000:01:00): 109, pid='<unknown>', name=<unknown>, Ch 0000009e, errorString CTX SWITCH TIMEOUT, Info 0x54c050
1 Like

Tested Elden Ring again with the latest drivers since the new DLC came out, and still get crashes in anywhere between 5 and 15 minutes of gameplay. Completely unplayable, not even in short sessions. Tested stable, beta and Vulkan dev drivers: 550.90.07, 555.52.04 and 550.40.63.

Also confirmed that, on my setup and with this game, crashes happen even right after driver updates.

Running an RTX 2060 Mobile on Arch Linux with standard kernel 6.9.5.

nvidia-bug-report-eldenring-550.90.07.log.gz (929.9 KB)
nvidia-bug-report-eldenring-555.52.04.log.gz (979.7 KB)
nvidia-bug-report-eldenring-550.40.63-vulkan.log.gz (916.5 KB)

Souns like shader cache handling problem

are you using gamescope? if you do, you need to disable swapchain extension

Unfortunately, I’m not. Might try using it and tinkering with the configs to see if it impacts the problem in any way, though.

555.58 released today, problem persists. Tested with the same game.

journalctl-eldenring-555.58.txt (224.3 KB)
nvidia-bug-report-eldenring-555.58.log.gz (1.1 MB)

Two more crashing applications on my RTX 2060 Mobile: Risk of Rain 2 and Final Fantasy XIV. Both games worked fine a couple months ago. At this point I’m unsure if my graphics card is faulty or what might be causing the problem besides lack of support for this issue. Will start testing different distros this week to make sure I don’t have a botched Linux install. This is incredibly frustrating.

nvidia-bug-report-ffxiv-555.58.log.gz (1.1 MB)
nvidia-bug-report-riskofrain2-555.58.log.gz (1.1 MB)

I’ve encountered the dreaded xid 109 system freeze on multiple games, from BG3 and The First Descendant to Disciples 2 (yep, even that one). Independent of proton version or tinkering steps, both BG3 and FD first ran fine for more than an hour, then the day after they started freezing after ~5-10 mins of gameplay. Elden Ring ran fine for like an hour on first day too, haven’t tested after.
nvidia-bug-report.log.gz (413.3 KB)

Tested Elden Ring. Ran with zero issues for ~5h yesterday, however, today it freezes after ~10 minutes of gameplay. You can see the xid 109 error in the logs.
nvidia-bug-report.log.gz (450.6 KB)

Tested on Helldivers 2, Ghost of Tsushima, and CS2. HD2 and GoT end up crashing with Xid 109, with seemingly no fixed time after which they crash - sometimes it’s 30 mins, sometimes it’s 10 mins. CS2 runs without any issues, tested on matchmaking, deathmatch, and casual.
Crash seems to occur both with and without “undervolting” (+core offset and a clock limit).

Here are my specs:
Tumbleweed 20240711, running KDE + Wayland
Linux v6.9.7-1-default
AMD Ryzen 7 5800X3D
RTX 3080 FE 10G (nvidia driver 550.90.07)
32GiB DDR4 RAM

You may update your Tumbleweed. Kernel 6.9.9 and nvidia 550.100 should already be installed. Not that I expect it to fix the issue but since you’re running an older GPU driver it might be worth a shot.

Then re-try and also post nvidia-bug-report.tar.gz obtained with sudo nvidia-bug-report.sh.

Thanks for the reply man. As you expected, even with the updated stuff the issue still persists (at least on Ghost of Tsushima)

nvidia-bug-report.log.gz (1015.0 KB)

Issue still happens. Was playing Wasteland 3:

[ 5949.384282] [   T1800] NVRM: GPU at PCI:0000:01:00: GPU-d94a8dd3-7c3d-ecc0-033e-1ec177ee9969
[ 5949.384286] [   T1800] NVRM: Xid (PCI:0000:01:00): 109, pid='<unknown>', name=<unknown>, Ch 0000001e, errorString CTX SWITCH TIMEOUT, Info 0x2c00e

nvidia-bug-report.log.gz (962.2 KB)

Getting freezes in Mechwarrior 5

version 550.78

kernel: NVRM: Xid (PCI:0000:08:00): 109, pid=‘’, name=, Ch 00000106, errorString CTX SWITCH TIMEOUT, Info 0x80c08f

Hi @mathewtodd84 @CommonOddity
Could you please try with recent 560 release driver and post test results.

Hello, the same problem (error 109) with pytorch - it’s working on cloud server, but not on my local machine. I was tested with Driver Version: 560.35.03, NVIDIA GeForce RTX 4090

NVRM: Xid (PCI:0000:01:00): 109, pid='<unknown>', name=<unknown>, Ch 0000000a, errorString CTX SWITCH TIMEOUT, Info 0x3c008

Thu Aug 29 13:10:43 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.35.03              Driver Version: 560.35.03      CUDA Version: 12.6     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 4090        Off |   00000000:01:00.0 Off |                  Off |
|  0%   23C    P8             20W /  450W |      15MiB /  24564MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A      1749      G   /usr/lib/xorg/Xorg                              4MiB |
+-----------------------------------------------------------------------------------------+

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Fri_Jun_14_16:34:21_PDT_2024
Cuda compilation tools, release 12.6, V12.6.20
Build cuda_12.6.r12.6/compiler.34431801_0

# gpu    pwr  gtemp  mtemp     sm    mem    enc    dec    jpg    ofa   mclk   pclk
# Idx      W      C      C      %      %      %      %      %      %    MHz    MHz
    0     21     22      -      0      0      0      0      0      0    405    210
    0     21     22      -      0      0      0      0      0      0    405    210
    0     21     22      -      0      0      0      0      0      0    405    210
    0     21     22      -      0      0      0      0      0      0    405    210
    0     21     22      -      0      0      0      0      0      0    405    210
    0     22     22      -      0      0      0      0      0      0    405    210
    0     21     22      -      0      0      0      0      0      0    405    210
    0     84     32      -     73     54      0      0      0      0  10251   2730
    0    224     27      -    100      0      0      0      0      0  10251   2775
    0     85     26      -    100      0      0      0      0      0  10251   2775
    0     84     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     82     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     84     26      -    100      0      0      0      0      0  10251   2730
    0     84     26      -    100      0      0      0      0      0  10251   2730
    0     83     26      -    100      0      0      0      0      0  10251   2730
    0     74     26      -      0      0      0      0      0      0  10251   2730
    0     69     26      -      0      0      0      0      0      0  10251   2730
    0     63     25      -      0      0      0      0      0      0  10251   2520
    0     60     25      -      0      0      0      0      0      0  10251   2520

nvidia-bug-report.log.gz (440.8 KB)

I was bored and took another look at this, seems to be a race condition, if you wait a bit, maybe 10-15 seconds at the main menu before loading a save, the game seems to work. If you load a game too fast, it will crash with XID 109. I managed to recreate the same crash in windows too using vkd3d-proton. So, yeah… either it’s a general driver issue or it’s something in vkd3d-proton.

The last of us XID 109 crash is fixed in Vulkan Beta 550.40.71

the same problem (error 109) with onnxruntime-gpu
Driver Version: 550.107.02 CUDA Version: 12.4 RTX 4090