Issues arise when performance peaks during heavy input/output operations.
Issues arise when performance peaks during heavy input/output operations.
CPU (1600X → 3800XT → 5800X), motherboard (from Asus B350-F to B550-Creator), RAM (16GB → 32GB with XMP disabled), GPU (GTX770 → RX 5600 XT → RTX 3060 Ti → RX 6600), power supply (CX650M to RM750x), case (now Hyte Y60), storage (Samsung 960 PRO → 990 PRO), operating system (reinstalled multiple times, except after RAM swap). The cooling setup... [now holds no original parts] I've run various tests—memtests, swapped cards, used UPS—and the odd thing is my nearly identical machine still performs well. My next actions will include re-seating memory and reinstalling Windows (which was on the agenda since Win11). Are there any other steps I'm missing? Maybe swapping the 5800Xs or even contacting AMD for an RMA?
The issue may occur with monitors above 1080p, not just ultrawides. Since dump files aren't available, we should examine crash error logs. In Event Viewer, locate the Kernel-Power event ID 41 crashes—if it's 6008, no details are recorded. Go to the Details tab and capture 3-5 screenshots. You can right-click each and choose "Save As" (multiple selections work) to upload the .evtx file.
Here’s a revised version of your text:
I learned you can retrieve the BSoD code from error 41 entries. Initially, I thought these codes only signaled an unexpected shutdown without any details about the cause. Most recent ones are 0x124 with WHEA_UNCORRECTABLE_ERROR. I upgraded the BIOS a few months back (after noticing a 2021 BIOS still installed). It’s now likely set to version 3802. I reset the BIOS last night and restored everything from the beginning. (Why do fan profiles appear in the BIOS settings? Why is fan control so complicated?) I also refreshed the chipset drivers based on some talks about Zen3 and the 0x124 BSoDs. I tried a few fixes that previously led to problems, but nothing changed—though it wasn’t the same issue that was causing trouble before. I was in the kitchen, not the bedroom, and the HDD docks were there. I used y-cruncher 5B, which worked well. By chance, when it crashes, it removes my 10G switches (just those) from under my Unifi controller. I admit I have some unusual edge-case bugs. I kept crashing it three times before combining those two changes.
Spoiler - <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>Microsoft-Windows-Kernel-Power</System>
<Provider Name="Microsoft-Windows-Kernel-Power" Guid="{331c3b3a-2005-44c2-ac5e-77220c37d6b4}" Version=8 Level=1 Task=63 Opcode=0
<TimeCreated>2025-06-17T22:10:18.276535</TimeCreated>
<EventID>41</EventID>
<Keywords>0x8000400000000002</Keywords>
</System>
Spoiler - <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>Microsoft-Windows-Kernel-Power</System>
<Provider Name="Microsoft-Windows-Kernel-Power" Guid="{331c3b3a-2005-44c2-ac5e-77220c37d6b4}" Version=8 Level=1 Task=63 Opcode=0
<TimeCreated>2025-06-16T23:58:45.3227</TimeCreated>
<EventID>41</EventID>
<Data>BugcheckCode=292</Data>
<Parameter1>0x10</Parameter1>
<Parameter2>0x0</Parameter2>
<Parameter3>0x0</Parameter3>
<Parameter4>0</Parameter4>
<SleepInProgress>true</SleepInProgress>
<PowerButtonTimestamp>0</PowerButtonTimestamp>
<BootAppStatus>true</BootAppStatus>
<Checkpoint>true</Checkpoint>
<ConnectedStandbyInProgress>false</ConnectedStandbyInProgress>
<SystemSleepTransitionsToOn>true</SystemSleepTransitionsToOn>
<CsEntryScenarioInstanceId>0</CsEntryScenarioInstanceId>
<BugcheckInfoFromEFI>true</BugcheckInfoFromEFI>
</EventData>
The zero value isn't caused by a BSOD, meaning it's not a 0 error. So NVMe is likely the main concern. I'm unsure if you experienced this problem with your older SSD, but since storage errors are usually driver-related (other WHEA issues point to CPU), it might be the driver, M.2 slot, or motherboard.
It looks like a fascinating situation... I swapped both NVMe units—both are Samsung PRO models, so I’m not assuming they’re faulty right away. After further research, I increased the SoC voltage to 1.15V. The system stayed steady for over 1.2TB transfers at about 350MBps for several hours, so I’ll consider revisiting this if it happens again. Might Infinity Fabric instability show up as an NVMe issue? I’m curious whether just taking the device off AC helped reset something.
I checked that scenario before; it wasn’t an issue. When my GPU caused the problem, I rebuilt the whole system in the old case and still found nothing (the recently upgraded and replaced GPU worked just fine).