Question WHEA-Logger 18 only while playing a specific game

[Mattheww Raven]

PC Configuration:
CPU - AMD Ryzen 7 5700X3D
GPU - Radeon 6750XT (XFX Speedster QICK 319)
Motherboard - Gigabyte B450 AORUS ELITE V2
Memory - 2X Corsair Vengeance LPX (1 x 8GB) DDR4-2400 (CL 15 and CL 16)
PSU: Corsair RM750x SHIFT 750W 80+ Gold Modular
Case: Rise Mode X4 Glass Black
Operating System & Version: WINDOWS 11 PRO 24H2 / 26100
GPU Drivers: AMD Adrenalin 25.3.1
Chipset Drivers: AMD CHIPSET DRIVERS VERSION 7.02.13.148

Ever since Monster Hunter Wilds came out (and even before during both beta tests), it will randomly restart my PC. The screen goes black, everything freezes and the PC reboots, and the black screen and freezing noises(?) differs depending on how long i've been playing, taking longer to reboot, and spending more time on the BIOS startup screen if it's been around 2 hours. Speaking of time, I've been monitoring it, and it usually crashes around 20, 30, 40, 01:20, sometimes going as far 02:30. And before upgrading my PSU, it would sometimes crash after 5 to 10 minutes, sometimes just 1.

There's only one other game where the same issue happened (Marvel Rivals), and it only happened once in 280 hours.

I've found two notable errors in event viewer; WHEA-Logger 18 (32 errors), and OneCore-DeviceAssociationService 3503 (103 errors). The amount of errors is obviously due to all my testing, although OneCore-DeviceAssociationService seems to be more frequent, and I couldn't find much relevant info about it, or if and how it relates to the other one.

Prior to this issue, I've never touched the BIOS other than for enabling XMP profile, or did any form of overclock or undervolt. Also, the CPU, GPU and motherboard are all less than one year old.

WHEA-Logger 18
A fatal hardware error has occurred.
Reported by component: Processor Core
Error Source: Machine Check Exception
Error Type: Cache Hierarchy Error
Processor APIC ID: 7

The processor ID changes everytime, and I've made a log of how much each showed, if that's relevant I will share it.

OneCore-DeviceAssociationService ID 3503
The Device Association Service dedected an error discovering endpoints.

I've been trying all sorts of solutions, with no sucess, such as:

• Cleaning my PC due to potential heat issues. I've also monitored temps during some game sessions, and nothing seems out of the ordinary, except for maybe the GPU hotspot, which hovers around 105º, which seems to be normal for my model.
• Multiple SFC and DISC checks (some of which fixed stuff, like the bluetooth drivers, just not the main issue)
• Swapping energy plan between Balanced and High Performance.
• Adding the game executable to windows defender exceptions.
• Verifying the game files on Steam
• Disabling Frame Gen, and lowering graphical settings
• Updating BIOS to F68d (I don't recall which version it was before updating, but most likely it was F65)
• Installing the game in a diferent drive, from my NV2 1TB M.2, to my C:\ Drive, a A400 240GB
• Upgrading my PSU from a Corsair 550W Bronze, to a Corsair RM750x Gold.
• Updating drivers and windows, even doing a clean install and upgrading(?) to Windows 11, and using AMD's Pro driver instead of Adrenalin.
• Changing BIOS settings: disabling C-State Control, Core Performance Boost, PBO and PBO Scalar. Something worth noting, in my BIOS, there's two C-State and CPB settings, two of each were in the XFR Enhancement tab.
• Enabling and disabling X.M.P profile.
• With PBO2 Tuner core offset, I've also tried -30 and -20 on all cores, while also setting PPT, TDC and EDC limits to 100, 70, 100, respectively.
• Undervolting(?) the CPU trough BIOS by setting the Dynamic Vcore (DVID) to -1, and increasing VCORE SOC to 1.
• Undervolting the GPU trough AMD Adrenalin and PRO, with a setting of 2700MHz, 1150mV and max Power Limit.
• Running Windows Memory Diagnostic, and also testing my RAM Stick of 5 years individually, on a different slot.
• Running Windows Driver Verifier, which actually caused a BSOD on startup (SYSTEM_THREAD_EXCEPTION_NOT_HANDLED) "caused" by ntoskrnl.exe, which doesn't really tell anything, being a Windows process.
• Reseting BIOS with CMOS battery.
• Going for two 8 (6+2) separate cables for the GPU.
• I've also tried to increase the CPU voltage, but as far as I know there's no way of doing that on my 5700X3D. Or at least, I couldn't find any settings for the CPU cores, or any info online other than the fact that this CPU doesn't seem to allow that. Maybe PBO2 Tuner has that option? I've tried increasing all cores to 5, but it didn't seem to change anything on OCCT's stress test. I thought about raising it higher, but I don't want to risk permanent damage (I've already killed my RX 580 some years ago, and would rather not take any risks).

I think this is most likely a CPU issue, although i'm not knowledgeable enough to confirm. It could also simply be that the issue stems from the game's poor optimization, seeing how many people have had problems related to stablity. Still, I can't help but feel like this is most likely a hardware fault, especially considering the Driver Verifier test.

 

[Lutfij]

Welcome to the forums, newcomer!

Case: Rise Mode X4 Glass Black
https://www.lojarisemode.com.br/pro...1sPz6sHMww5hy76sih7uRv788HjvdpEeM0C1Xo97DpNuH
Is this ^ your case? Do you get better temps when you remove the front fascia altogether? Make and model of the fans in your case?

Updating BIOS to F68d (I don't recall which version it was before updating, but most likely it was F65)
Did you clear the CMOS after verifying your BIOS was successfully flashed to F68d? Ideally you're advised to disconnect from the wall and display, then remove the CMOS battery, press and hold down the power button for 30secs to drain any residual power, then reseat the battery after 30mins.


Updating drivers and windows, even doing a clean install and upgrading(?) to Windows 11
Did you migrate to Windows 11 using the internal upgrade path from Windows 10? If so, you're advised to reinstall Windows 11 after recreating your bootable USB installer for Windows 11, installing the OS in offline mode.

I've also tried to increase the CPU voltage, but as far as I know there's no way of doing that on my 5700X3D.
That's not how you should work with the Ryzen processors, you're advised to use Ryzen Master's Curve Optimizer and see if that nets you lower power draw/temps.

I think this is most likely a CPU issue, although i'm not knowledgeable enough to confirm.
In all of this, since your PCPartPicker link is privated and impossible to view, what are you using to cool your processor?

It could also simply be that the issue stems from the game's poor optimization, seeing how many people have had problems related to stablity.
Yes, that too.

Memory - 2X Corsair Vengeance LPX (1 x 8GB) DDR4-2400 (CL 15 and CL 16)
Please elaborate on this. Are you using two sticks of ram purchased at different times? If so, check to see their PCB revisions. Just an FYI, you're platform is losing out on performance. Your Ryzen platform would benefit most with a dual channel, tight latency DDR4-3200MHz ram kit.

 

[Mattheww Raven]

Thanks for responding.

Case: Rise Mode X4 Glass Black
https://www.lojarisemode.com.br/pro...1sPz6sHMww5hy76sih7uRv788HjvdpEeM0C1Xo97DpNuH
Is this ^ your case? Do you get better temps when you remove the front fascia altogether? Make and model of the fans in your case?

Yes, that's the correct case. I haven't done this temp check, but based on HWINFO's log, temps are normal. Besides, I've played other demanding games for much longer without issues, including Dragon's Dogma 2, another poorly optimized game that share's a lot of DNA with MH: Wilds (Same engine, open world, and other similar features).
As for the fans, I use two of these cheap ones (120mm, 7 blades, up to 1200 RPM and 36 CFM). One in the front, and the other in the back, with the cooler in the middle, all in the correct side for proper airflow.

Updating BIOS to F68d (I don't recall which version it was before updating, but most likely it was F65)
Did you clear the CMOS after verifying your BIOS was successfully flashed to F68d? Ideally you're advised to disconnect from the wall and display, then remove the CMOS battery, press and hold down the power button for 30secs to drain any residual power, then reseat the battery after 30mins.

I didn't clear CMOS immediately after, but I did check and it said F68d. When I did clear the CMOS, I just unplugged everything, turned off the PSU, removed it and left it out for around 2 minutes before putting it back. When I booted into the BIOS, it confirmed that the setting had been reset.
Something worth mentioning: after reading your post, I went to check the BIOS to double check the version, and it once again said that the configuration had been reset. I did the CMOS thing 2 days ago, and yesterday installed a new cooler, but didn't touch the CMOS. Maybe I didn't put it back properly?

Updating drivers and windows, even doing a clean install and upgrading(?) to Windows 11
Did you migrate to Windows 11 using the internal upgrade path from Windows 10? If so, you're advised to reinstall Windows 11 after recreating your bootable USB installer for Windows 11, installing the OS in offline mode.

I installed it from Windows 10 using a bootable USB installer, although it was online.

I've also tried to increase the CPU voltage, but as far as I know there's no way of doing that on my 5700X3D.
That's not how you should work with the Ryzen processors, you're advised to use Ryzen Master's Curve Optimizer and see if that nets you lower power draw/temps.

That's the thing, this particular model (5700X3D) doesn't have Curve Optimizer on the BIOS or Ryzen Master, PBO2 Tuner is the only way of undervolting. This lock is the same as the 5800X3D model.

I think this is most likely a CPU issue, although i'm not knowledgeable enough to confirm.
In all of this, since your PCPartPicker link is privated and impossible to view, what are you using to cool your processor?

Sorry about that, just made it public.
As for the cooler, I just got a new one yesterday. Went from a 5 year old Deepcool Gammaxx 300R to a Thermalright Assassin Spirit 120.

Memory - 2X Corsair Vengeance LPX (1 x 8GB) DDR4-2400 (CL 15 and CL 16)
Please elaborate on this. Are you using two sticks of ram purchased at different times? If so, check to see their PCB revisions. Just an FYI, you're platform is losing out on performance. Your Ryzen platform would benefit most with a dual channel, tight latency DDR4-3200MHz ram kit.

Yes, I have 2 RAMS. One I have since 2020, the other one I got a few months ago.
Just checked the PCB revisions and yes, they're different. The older one is v. 4.32, while the newer one is v. 5.38.

I'll be checking some new RAM, since maybe the lower clock speed and different PCB could explain the fact that this issue seems mostly exclusive to MH: Wilds. Also, as mentioned earlier, this happened once in Marvel Rivals, and I recall it being during the exact moment I used a specific ability, Dr. Strange's Portal, which does what you think it does, and is very notorious for how it eats everyone's framerate, especially during cast.

You think a Corsair LPX (2x16GB) 3600MHz would be a good upgrade? Pretty good deal it seems, considering the 3200 kits are currently more expensive (for 32GB dual channel).