My computer has been randomly crashing with a DPC_WATCHDOG_VIOLATION (0x133) blue screen, and the crashes have become more frequent this year. Before each crash, programs stop responding even though the mouse still moves; then the mouse freezes, followed shortly by the BSOD. Recently the crash screen has appeared black, but it still shows the same error code.
I initially suspected an outdated BIOS because my previous version was from 2021. After updating to the latest recommended version, the system became more stable, but it can still crash after sitting idle for a while. Interestingly, it usually stays stable when a YouTube video or a game is running. One exception happened while playing GTA V with a mod menu enabled, when the console displayed an exception before the same watchdog error appeared.
I stress-tested both the CPU and GPU with FurMark without crashes or graphical artifacts, so I'm unsure whether either component is failing. I'm looking for suggestions on how to narrow this down, especially whether XMP, memory, storage, drivers, or the power supply could be involved. Checking the Windows minidump files may also help identify the failing driver or component.
System specifications:
- Ryzen 9 5950X
- GeForce RTX 3090 with Game Ready drivers
- 32 GB Corsair Vengeance DDR4 RAM, 4 x 8 GB, XMP enabled at 3200 MHz
- MSI MPG B550 Gaming Plus motherboard
- Gigabyte P750GM V2, 750W, 80 Plus power supply
- Crucial P3 1 TB NVMe SSD
- Samsung 870 QVO 1 TB SATA SSD
- Seagate Constellation 4 TB SATA drive
- TP-Link TL-WN881ND Wi-Fi card
3 Answers
The power supply is worth taking seriously here. A 5950X paired with an RTX 3090 can have substantial transient power demand, and a 750W Gigabyte P750GM is not an ideal match for that hardware. Replacing it with a reputable, high-quality 850W or larger unit could eliminate power-related instability, particularly if the system behaves differently during idle power-state changes.
First, check whether the crashes continue with XMP disabled. Four memory modules running with an overclocked memory profile can sometimes expose stability problems, even if the system passes normal CPU and GPU stress tests. Run a memory test as well, and if possible test the modules individually or with fewer sticks installed. The minidump files in C:WindowsMinidump would also be important for identifying the actual driver involved.
Stress testing does not rule out memory, storage, Wi-Fi, or driver problems. Since this is a watchdog error, collect several minidumps rather than guessing from the error name alone. Also check storage health, update motherboard and chipset drivers, temporarily remove the Wi-Fi card or its driver, and test with XMP off. The fact that activity seems to prevent the crash may point to an idle power-state or device-driver transition rather than a simple overheating issue.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures