My ASUS ROG Strix G15 G513IH (Ryzen 7 4800H, GTX 1650 Laptop GPU, 16 GB RAM, BIOS 316) has been experiencing inconsistent BSODs, including KMODE_EXCEPTION_NOT_HANDLED (0x1E), IRQL_NOT_LESS_OR_EQUAL (0xA), and DRIVER_OVERRAN_STACK_BUFFER (0xF7). Several WinDbg reports show badly corrupted or invalid stack data, so I'm unsure whether the named driver is actually responsible or whether corruption occurred earlier.
One recurring lead is nvpcf.sys, along with nvlddmkm.sys, AMD power-management drivers such as amdppm.sys and AmdMicroPEP.sys, and ASUS system components. The !blackboxpnp data also shows PnpProblemCode 22 for ACPI\NVDA0820\NPCF. I temporarily disabled nvpcf and went about a day without a crash, but I don't consider that conclusive.
The most recent crash occurred while playing Warhammer 40,000: Space Marine 2. It reported KMODE_EXCEPTION_NOT_HANDLED with Arg1 ffffffff80000003, and the failure bucket referenced nt!KeGetCurrentStackPointer. The reported instruction was int 3, followed by a heavily corrupted-looking stack.
I have already completely removed and reinstalled both NVIDIA and AMD drivers with DDU, and I have tested the RAM. I'm trying to determine whether nvpcf.sys and the ACPI problem code are meaningful, whether they could cause this level of stack corruption, or whether the real cause is more likely mixed memory, an AMD chipset or power-management issue, BIOS/EC firmware, hardware instability, or something else. What would you investigate next, and are there any particularly stable driver versions for this laptop configuration?
3 Answers
STATUS_BREAKPOINT (0x80000003) at an unexpected location is not enough to conclude that the CPU is defective. With corrupted control flow, it can be the result of earlier memory corruption, a faulty driver, unstable RAM timings, firmware, or another hardware problem. I would collect several small kernel or minidumps, enable Driver Verifier selectively for non-Microsoft drivers, and test stability with GPU acceleration and power-management features changed one at a time. Avoid enabling every verifier option at once, since that can make the system unbootable and obscure the diagnosis.
The ACPI\NVDA0820\NPCF entry and Problem Code 22 indicate that the NVIDIA-related ACPI device is disabled or not currently started. That makes nvpcf.sys worth investigating, but it is not conclusive evidence that it caused the crash. A device being present in loaded-module data or near a crash can simply mean Windows detected a problem there. Compare crashes with the component disabled, and check the ASUS BIOS, chipset, firmware, and NVIDIA driver packages rather than focusing only on the dump’s blamed module.
Mixed bugchecks combined with badly corrupted stacks usually point to memory or broader hardware instability rather than proving that nvpcf.sys is defective. Even if the RAM passed a quick test, run a bootable test such as MemTest86 for several full passes or overnight. Also reseat the modules and check whether they are matched modules; identical advertised speed does not guarantee identical timings or memory behavior.
The BIOS is already current, and I have removed and reinstalled the graphics and AMD drivers. The laptop has one Micron and one Kingston module with matching specifications, although they are not from the same kit.