dxvk: Monster Hunter World randomly freezes

Monster Hunter World (with proton) randomly freezes. This usually happens in between after 10min to 4hours, so a long random time period.

As the DXVK_HUD (with memory) was enabled, at the time of the freeze, around ~3.9gb (assuming this is vram) of 6gb were used.

Most noticeable the dmesg output:

NVRM: Xid (PCI:0000:09:00): 31, Ch 0000004b, intr 10000000. MMU Fault: ENGINE GRAPHICS GPCCLIENT_T1_4 faulted @ 0x0_00000000. Fault is of type FAULT_PDE ACCESS_TYPE_READ

Xid 31, the addr 0x0_00000000, intr 10000000 and ACCESS_TYPE_READ are always constant.

To me, it looks like a simple nullptr access, as it is always the 0x0 addr, but I don’t know how to investigate this problem further. I can not let the game run with VK_INSTANCE_LAYERS=VK_LAYER_LUNARG_standard_validation or apitrace for hours, as this makes it very unplayable.

PROTON_USE_WINED3D=1 just results in a black screen. Allow flipping (in nvidia-setting) on/off does not change anything.

Please let me know how to make this report more useful, I am out of ideas.

Software information

Monster Hunter World
vsync: off
30fps lock (getting weird input lag otherwise sometimes)
Steam / Proton 3.16-beta5

System information

GPU: Nvidia Geforce 1060gtx 6gb
Driver: nvidia-drivers-415.23
Wine version: Proton 3.16-beta5 (???)
DXVK version: Proton 3.16-beta5 (dxvk 0.93)
Kernel: 4.19.10
Ram: 16gb
CPU: Ryzen 2700X

Log files

(with DXVK_LOG_LEVEL=debug and DXVK_HUD=devinfo,fps,memory)

d3d11.log: MonsterHunterWorld_d3d11.log
dxgi.log: MonsterHunterWorld_dxgi.log

EDIT: The game overall runs pretty well, just the random freezes are a pretty frustrating problem.

EDIT2: The screen freezes but the game background music is still running.

About this issue

Original URL
State: closed
Created 6 years ago
Comments: 79 (8 by maintainers)

Most upvoted comments

Reached out to Nvidia to see if they can help debug it

Xaenalt on Jun 17, 2019

Tested the latest vulkan beta drivers (435.19.03) and the issue persists (freeze after ~3h 50m), so this wasn’t fixed by the fix for #1169 unfortunately.

I was running nvidia-smi dmon -s pucm -o DT -i 0 during the testing and this was towards the end of the output, fifth line being where the freeze occurred:

 20190909   16:00:15      0   131    63     -    99    47     0     0  4006  1936  3191     9
 20190909   16:00:16      0   132    63     -    99    47     0     0  4006  1936  3191     9
 20190909   16:00:17      0   131    63     -    99    47     0     0  4006  1936  3191     9
 20190909   16:00:18      0   133    63     -    99    47     0     0  4006  1936  3191     9
 20190909   16:00:19      0    49    62     -     0     0     0     0  4006  1936  3191     9
 20190909   16:00:20      0    49    61     -     0     0     0     0  4006  1936  3191     9
 20190909   16:00:21      0    49    61     -     0     0     0     0  4006  1936  3191     9
 20190909   16:00:22      0    33    60     -     0     1     0     0  3802   999  3191     9
#Date       Time        gpu   pwr gtemp mtemp    sm   mem   enc   dec  mclk  pclk    fb  bar1
#YYYYMMDD   HH:MM:SS    Idx     W     C     C     %     %     %     %   MHz   MHz    MB    MB
 20190909   16:00:23      0    33    59     -     0     1     0     0  3802   999  3191     9
 20190909   16:00:24      0    24    59     -     0     2     0     0   810   797  3191     9
 20190909   16:00:25      0    17    58     -     0     3     0     0   810   797  3191     9
 20190909   16:00:26      0    17    58     -     0     3     0     0   810   797  3191     9
 20190909   16:00:27      0    15    57     -     1     5     0     0   405   227  3191     9
 20190909   16:00:28      0    15    57     -     1     6     0     0   405   227  3191     9

At the start, the line listed sm % as 99 and mem % as 46. Normal usage shows sm % of 0 or 1, and mem % of 1.

Game window froze, but terminal on other monitor continued to output every second up until I tried to alt+tab out, then everything froze. Used an active SSH session to killall -9 MonsterHunterWorld.exe to kill MHW and everything unfroze, with no need to kill Xorg.

dmesg shows the same Xid error:

[2247625.011437] NVRM: GPU at PCI:0000:01:00: GPU-1ec9083a-db9c-1b2f-fd59-87c82dd1c09a
[2247625.011440] NVRM: GPU Board Serial Number: 
[2247625.011442] NVRM: Xid (PCI:0000:01:00): 31, pid=11036, Ch 0000003b, intr 10000000. MMU Fault: ENGINE GRAPHICS GPCCLIENT_T1_5 faulted @ 0x0_00000000. Fault is of type FAULT_PDE ACCESS_TYPE_READ

If an API Trace or something else can help with diagnosing the cause of this issue, I can generate one provided someone can provide me documentation on how to do so. I have no issue letting the game run overnight to get a valid dump if needed.

valarnin on Sep 9, 2019

Same result with 418.42.02. Started at 1551910780.835805, froze at 1551931937.174271, just under 6 hours of runtime.

valarnin on Mar 7, 2019