Need More Info HARD Crash on save 1.8 and 1.9

[Apo77o]

Summary: I have tried both 1.8 main and 1.9 beta branch and this crash happens on both. I have clean reinstalled 2x. updated all drivers and software, factory reset my overclocks, tested all hardware, tried with mods on and off, uninstalled mod, reinstalled mod (only using point per level mod) I tried starting multiple sandbox campaigns all experienced the same failure. Ran hardware stress tests and heat/power is fine. Secuity settings are all open. (I have over 241 hours in game prior to 1.7 when I took a break so its a new bug)

How to Reproduce: Save the game. When starting a new campaign can save 3-4 times then crashes the computer. After that it will crash on every save attempt autosave/save as/save
Have you used cheats and if so which: no
Scene Name (if related): Any
Media (Screenshots & Video): Black screen, peripherals off, LEDs on
Computer Specs:
OS: Windows 10
GPU: ASUS Strix 3090
GPU Driver Version: 517.48
CPU: AMD 5950X
RAM: G.Skill 128GB
Motherboard: ASUS Dark Hero
Storage Device (HDD/SSD): WD SN 850 NVME (also tried on Samsung 870 evo as both C and E drives)
PSU: EVGA 1000W gold
Monitor: Zowie XL 2546k 1080p

 

[MArdA TaleWorlds]

What is your PSU? Could you check if your PSU is well enough for your rig? The only explanation we thought is that the game uses multi-core when saving happens and that might be triggering it. Which leads to a CPU that is consuming more power.

 

[Apo77o]

What is your PSU? Could you check if your PSU is well enough for your rig? The only explanation we thought is that the game uses multi-core when saving happens and that might be triggering it. Which leads to a CPU that is consuming more power.

EVGA 1000W gold "recommended" is 642W. Which should be more than enough. All of the ECO/power saving features are off, and it's configured to give max power. I don't have a bigger PSU on hand, but I have run pretty serious overclocks on different benchmarks to see if it was a hardware issue and I haven't found anything. I will try to get a hold of a 1600w platinum PSU to see if I can rule it out altogether.

I have another pc with a very similar setup I'm going to try testing on that and see if I can figure out what is going on. This issue has me stumped so I'm open to any ideas. The last 4 test games I have run have this start on the 5th save. Creating a new save each time or overwriting each time didn't make a difference. I was playing through the game though. When I do my next tests I'll try just starting the game and repeat saving while the game is paused immediately after starting to see if that triggers it.

Also new video driver from NVidia today version 522.25 will begin testing on that.

 

[Apo77o]

Update: Game crashes on the 3-4 save if overwriting the same save. 4-5 save if create new save each time. Saving while stationary, moving, in town, out of town doesn't make a difference. Windows Event Viewer doesn't show anything. tried uninstalling and reinstalling drivers no change. Ran 20 loops of 3d mark time spy stress test without any issues. I have run every test I can think of and everything is fine. I really have no idea what could be causing this crash.

 

[Sciohan]

Can you check the status of OneDrive, make sure it didn't get signed out and isn't stuck trying to sync?

Can you post the rgl_log and launcher_log to a pastebin or similar?

What's the general storage setup look like? Is C: on an SSD or HDD, through SATA/IDE/M.2/other? Is it connecting directly to the integrated chipset or through an internal or external adapter? Does %USERPROFILE% reside on C: or is it on another partition? What drives and interfaces are Steam and Bannerlord installed to? Is everything on a standard NTFS partition?

There really shouldn't be anything the game is capable of doing that would cause such a severe malfunction. There are however waste issues and the game will happily run your CPU/GPU into temperatures much higher than most games even if idling. You may want to try using a hardware monitor of choice to observe the voltages your MB reports going to your CPU / RAM / GPU / Chipset and see if any of them vary significantly while playing; even if you have a very good PSU, that doesn't necessarily mean that every power rail is flawless or that the motherboard itself doesn't have supply issues under load.

 

[Apo77o]

Can you check the status of OneDrive, make sure it didn't get signed out and isn't stuck trying to sync?

Can you post the rgl_log and launcher_log to a pastebin or similar?

What's the general storage setup look like? Is C: on an SSD or HDD, through SATA/IDE/M.2/other? Is it connecting directly to the integrated chipset or through an internal or external adapter? Does %USERPROFILE% reside on C: or is it on another partition? What drives and interfaces are Steam and Bannerlord installed to? Is everything on a standard NTFS partition?

There really shouldn't be anything the game is capable of doing that would cause such a severe malfunction. There are however waste issues and the game will happily run your CPU/GPU into temperatures much higher than most games even if idling. You may want to try using a hardware monitor of choice to observe the voltages your MB reports going to your CPU / RAM / GPU / Chipset and see if any of them vary significantly while playing; even if you have a very good PSU, that doesn't necessarily mean that every power rail is flawless or that the motherboard itself doesn't have supply issues under load.

OneDrive is off.
C is M.2 SMART drive remaining life is 99%
I have tried moving bannerlord to an SSD connected via sata
NTFS windows is on C so is steam and bannerlord
Cooling is a custom loop from EK 2x 360mm rads 1x 420mm rad 1 x 140mm rad. Fans are Noctua industrials. The 3090 is also setup with an active backplate for cooling the memory on the "back"
HWINFO GPUZ and Ryzen master all look fine before the crash Temps and Voltages.

Where are those logs at?

I'm getting to the point where I am hoping it's a hardware problem then it would be something I could fix. Which was my first thought when I started having this problem since I haven't been able to find a single report of something similar happening to anyone else. But every other game/software/test says the hardware is fine. Should be able to get computer on the bench tomorrow just pull the whole thing apart and test each component one by one.

 

[Sciohan]

OneDrive is off.
C is M.2 SMART drive remaining life is 99%
I have tried moving bannerlord to an SSD connected via sata
NTFS windows is on C so is steam and bannerlord
Cooling is a custom loop from EK 2x 360mm rads 1x 420mm rad 1 x 140mm rad. Fans are Noctua industrials. The 3090 is also setup with an active backplate for cooling the memory on the "back"
HWINFO GPUZ and Ryzen master all look fine before the crash Temps and Voltages.

Where are those logs at?

I'm getting to the point where I am hoping it's a hardware problem then it would be something I could fix. Which was my first thought when I started having this problem since I haven't been able to find a single report of something similar happening to anyone else. But every other game/software/test says the hardware is fine. Should be able to get computer on the bench tomorrow just pull the whole thing apart and test each component one by one.

The logs for your most recent session should be stored in C:\ProgramData\Mount and Blade II Bannerlord\logs and logs from uploaded crash reports will be in C:\ProgramData\Mount and Blade II Bannerlord\crashes

How is Windows crashing? Bluescreen, shutting off, other? Have you already tried closing as many background applications as you can? Any interesting entries in Windows Event Viewer on boot or around the time of the crash? I see you've already updated all drivers, but was that done through Windows Update or through ASUS' site? BIOS also up to date?

 

[Taheen19]

Summary: I have tried both 1.8 main and 1.9 beta branch and this crash happens on both. I have clean reinstalled 2x. updated all drivers and software, factory reset my overclocks, tested all hardware, tried with mods on and off, uninstalled mod, reinstalled mod (only using point per level mod) I tried starting multiple sandbox campaigns all experienced the same failure. Ran hardware stress tests and heat/power is fine. Secuity settings are all open. (I have over 241 hours in game prior to 1.7 when I took a break so its a new bug)

How to Reproduce: Save the game. When starting a new campaign can save 3-4 times then crashes the computer. After that it will crash on every save attempt autosave/save as/save
Have you used cheats and if so which: no
Scene Name (if related): Any
Media (Screenshots & Video): Black screen, peripherals off, LEDs on
Computer Specs:
OS: Windows 10
GPU: ASUS Strix 3090
GPU Driver Version: 517.48
CPU: AMD 5950X
RAM: G.Skill 128GB
Motherboard: ASUS Dark Hero
Storage Device (HDD/SSD): WD SN 850 NVME (also tried on Samsung 870 evo as both C and E drives)
PSU: EVGA 1000W gold
Monitor: Zowie XL 2546k 1080p

I believe I may know the problem and have a solution - though further testing is required.

Basically I have the exact same issue, and PC with eerily similar specs.

Summary: PC hard crashes (essentially turns off) when saving the game.

Computer Specs:
OS: Windows 10
GPU: Sapphire Toxic 6900XT
GPU Driver Version: 22.5.1
CPU: AMD 5950X
RAM: G.Skill F-4 3600C16 16GTZN 64GB
Motherboard: ASUS ROG CROSSHAIR VIII HERO (BIOS 4201)
Storage Device (HDD/SSD): WDC WDS100T2B0C-00PXH0 (also tried multiple drives)
PSU: EVGA 1000W G2 Supernova
Monitor: Samsung Odyssey G9 5120x1440 @120Hz (same issue at 60Hz and 240Hz)


What I discovered is this:

The game has high residency on Core 0 and Core 3

When saving, one of these two cores specifically (it seems to alternate) experiences a large power (and thus temp) spike. I believe this spike causes that core temp to very quickly pass the maximum safe junction temp for the core and trips the board to shutdown to prevent damage to the CPU.

Like OP I was initially running the CPU overclocked (via PBO2 and curve optimization) while also running high frame rates (between 200Hz and 240Hz). The PC would hard crash after every save.


I found that capping frame rates to 120Hz helped - PC would crash maybe 1 out of every 4 saves. I began to investigate CPU using HWinfo to log data. This is how I noticed temp spike when saving.


I have since completely disabled PBO in the mobo's overclocking tab and also disabled curve optimization. Also - and very important! - on my mobo when setting PBO limits (PPT, TDC, EDC) to Auto allows limits beyond the stock values (142, 95, 140, respectively for a 5950x) manually setting defaults or setting to "Motherboard" is the only way to ensure default PBO limits.


Since doing this I have not had a crash when saving the game - though I do still observe Hot Spot temp spikes into the low 80C's when saving.

It's only been a few days, and I'll report back to this thread if anything changes or crashes resume. I do still have some log files if devs would like a look - I'm confident you could reproduce this with a CrosshairVIII board, a 5950x, and the correct settings (PBO On, Curve optimizer On, and PBO limits to Auto or manually OC'd - mine were at PPT 185, TDC 130, EDC 170).


Hope this helps!