GPU: XFX Radeon RX 5700 XT THICC III (65-70 C on load)
CPU: Ryzen 5 3600x 6 CORE 12 THREADS (70 C on load)
COOLER: Arctic Liquid Freezer 2 240
Motherboard: MSI TOMAHAWK MAX B450
BIOS Version: 7C02v3J
RAM: 16GB G.Skill Trident Z RGB DDR4 3600 (F4-3600C16D-16GTZRC)
PSU: Corsair RM750 — 750 Watt 80 PLUS® Gold
Case: Cooler Master Mastercase 5
SSD: 970 Evo Plus + 870 Evo
Operating System & Version: Windows 10 Pro Version 22H2 - 19045.5608
GPU Drivers: Adrenalin 25.3.1 (WHQL Recommended)
Chipset Drivers: AMD B450 CHIPSET DRIVERS VERSION 7.2.13.148
Background Applications: BRAVE, EDGE
I built this system in 2020 and have been using it ever since. The first issue I encountered was in the early years, while playing Battlefield 1. The game would suddenly close, and I’d be returned to the desktop with a message that the AMD driver encountered an error. I don’t remember the exact error message, but it mentioned som
ething about the GPU being physically disconnected (?). I couldn’t resolve the issue, so I stopped playing the game.
My current issue has become more complex and frustrating.
Description of Original Problem:
Non-BSOD RESTART with WHEA ERROR ID 18 or Kernel 41
Situations That Trigger the Symptom:
Playing any game (The timing is random; it can happen instantly or within 1.5 hours. This happens especially with Battlefield 1 and Dark and Darker).
Switching to the desktop while playing a game (ALT+TAB)
Having YouTube running in the background while gaming.
Troubleshooting:
Increased and decreased CPU voltage (offset, override, AMD override).
Increased and decreased NB/SOC voltage (offset, override, AMD override).
Enabled and adjusted PBO.
Manually set CPB values and disabled it.
Used RAM profiles and increased voltage (offset, override, AMD override), and also set FCLK/MEMCLK values manually while keeping other settings on auto.
Tried all C-State options.
Reset all BIOS settings.
Rolled back and updated the BIOS (Interestingly, when I used version 7C02v30, the WHEA Cache Hierarchy error turned into Kernel 41, but the behavior remained the same).
Tested a spare 7700K CPU on a Maximus Hero VIII motherboard to rule out a faulty processor.
Tested two spare RAM kits (CMK16GX4M2B3000C15R and CMK16GX4M2B3200C16) to rule out faulty RAM.
Renewed the thermal paste of the GPU and confirmed temperatures remained below 70°C.
Changed the PCIe slot of the GPU to eliminate potential GPU and M.2 SSD conflicts.
Installed the system on a SATA SSD.
Adjusted video settings and fullscreen/borderless fullscreen
Additional Testing:
During a stress test using the latest Radeon Software, the PC restarted when I tried to take a screenshot. Here’s what I noticed during the stress test: GPU clock, VRAM, and power consumption intermittently dropped to zero. I couldn’t determine why this was happening.
The system restarted while I was taking this screenshot too. The test was still in progress flawlessly.
I suspect that during such drops in gaming, there might be a short circuit or a similar issue that causes the PC to restart.
This theory might be entirely incorrect, as I’m not an expert in computer systems. What do you think could be causing this issue? Has anyone experienced a similar situation and managed to resolve it without replacing the GPU or PSU?
Thanks in advance to everyone.
