The symptoms point first to the NVIDIA graphics path, and the context supports two likely buckets: a failing video driver path or failing graphics hardware.
VIDEO_MEMORY_MANAGEMENT_INTERNAL is usually caused by a video driver behaving improperly. For that stop code, the documented next step is to check Windows Update for an updated video driver. For VIDEO_TDR_FAILURE and related TDR issues, the documented causes also include defective parts, insufficient system power, insufficient cooling, incorrect component settings, and background load that prevents the GPU from responding in time.
Given the details already provided:
- the NVIDIA driver has already been removed and reinstalled many times
-
nvlddmkm keeps appearing
- the system shows graphical glitches and black screens
- the GPU crashes almost immediately in a GPU stress test even after heavy underclocking
The strongest next tests are hardware-isolation checks rather than repeating driver reinstall steps.
- Check Event Viewer for matching system errors around each crash.
- Open Event Viewer and review the System log.
- Look for critical or error events in the same time window as the crash.
- The goal is to see whether the failures consistently point to the display stack or also show power, PCIe, or memory-related faults.
- Run Windows Memory Diagnostics.
- In Control Panel, search for Memory.
- Select Diagnose your computer's memory problems.
- After the test completes, open Event Viewer → System log and look for MemoryDiagnostics-Results.
- This helps rule out system memory corruption contributing to video memory manager crashes.
- Boot into Safe Mode to isolate the issue.
- Go to Settings → Update and Security → Recovery.
- Under Advanced startup, restart.
- Choose Troubleshoot → Advanced Options → Startup Settings → Restart.
- After restart, choose option 4, 5, or 6 for Safe Mode.
- If the system is stable in Safe Mode, that supports a graphics driver or full-performance GPU path issue rather than a general Windows stability problem.
- Verify graphics software is current.
- Make sure graphics-related software such as DirectX and OpenGL dependent applications are fully up to date.
- Also ensure any graphics-intensive applications are fully patched.
- Run the laptop manufacturer’s hardware diagnostics.
- The documented guidance is to use the hardware diagnostics supplied by the system manufacturer.
- This is the best next step to check for broader hardware faults outside the driver stack.
- Check for recent hardware or configuration factors that can affect TDR behavior.
- Confirm there are no overclocked components or altered memory settings.
- TDR failures are documented to occur with insufficient cooling, insufficient system power, incorrect compatibility/settings, and defective parts.
Based on the pattern described, the evidence is more consistent with a hardware problem in the graphics subsystem than with a simple driver issue. That aligns with similar accepted cases where persistent nvlddmkm crashes, artifacts, and failures under GPU load pointed to the GPU as the likely failing component. On a laptop, that can mean the GPU itself or related motherboard power/PCIe circuitry.
A practical conclusion from the supported guidance:
- If Memory Diagnostics is clean,
- the system is unstable only when the NVIDIA path is active or under GPU load,
- and manufacturer diagnostics do not identify another component,
then the next likely fault domain is the laptop graphics hardware path rather than Windows configuration.
References: