Question Random Freezes / Black Screen / Auto Restart with nvlddmkm Errors (Event ID 14 & 153) – Assistance Required
Question Random Freezes / Black Screen / Auto Restart with nvlddmkm Errors (Event ID 14 & 153) – Assistance Required
My computer suddenly stops working, even when nothing demanding is running (like browsing or streaming). The display turns black occasionally, though audio might still play, and after about 2–3 minutes it restarts itself. There are no crash dumps in C:\Windows\Minidumps, so the only records I have come from Event Viewer and Reliability History.
The unusual aspect is that these freezes happen unpredictably—sometimes not for more than a day, even when the machine is idle. They don’t occur during gaming or high stress situations. The most effective way to mimic the issue is simply leaving the PC on until it freezes.
Recent Nvidia-related errors were observed:
nvlddmkm (Event ID 14)
nvlddmkm (Event ID 153)
After extensive research on Google and Reddit, trying all common solutions, the problem remains unresolved.
PC specifications:
- Built: January 2023
- GPU upgrade: March 2025 (switched from 1080 Ti to Gigabyte Nvidia 5080 OC)
- Worked well for about 4 months
- Began crashes in July 2025
Previously, I experienced a Chrome crash on i9-13900K, which was resolved by adjusting Intel XTU multipliers (see Reddit thread). However, that never led to black screens or restarts.
Around July 2025, the main issue started after updating BIOS from 2703 to 3001 (Asus Z790 Hero). This version introduced numerous BSODs, so I reverted to BIOS 2801, which was reported as more stable.
I rolled back the Intel chipset and ME drivers to match BIOS 3001. The BSODs ceased, but the freezing, black screen, and restart behavior continued.
What I attempted:
- Stress tested the PSU with HWiNFO (15–20 minutes) – no issues
- Memtest64 (BIOS) – passed
- Windows Memory Diagnostic – passed
- Reseated GPU & RAM
- Disabled XMP in BIOS
- Used NVIDIA Control Panel, set Power management to Prefer Maximum Performance
- Disabled hardware acceleration in BIOS
- Removed iCUE (some posts suggested it caused crashes)
- Performed DDU clean install of Nvidia drivers (not done during initial upgrade)
- Toggled settings before and after DDU
After these steps:
- Installed NVIDIA 572.83 (stable from March to July)
- Used NVIDIA App with Custom(Advanced) installation
- Removed primary monitor from GPU and reinserted it
- Disabled Fast Boot in BIOS
- Disabled Fast Startup in Windows
- Set Primary Display in BIOS (disabling Integrated Graphics)
- Enabled Error Reporting (ERP) in BIOS
- Disabled NVIDIA related audio devices in Device Manager
- Changed HDMI/DP cable to 120Hz on Samsung Odyssey G80SD
- Turned off Hardware Acceleration in Discord
Potential causes suggested by others:
- Faulty RAM stick
- PSU problems
- GPU 12VHPWR cable (included with GPU)
- Defective GPU (though it worked for months)
- CPU issues
- iCUE or RGB software conflicts
- Motherboard fault (rare)
- Even after a clean Windows install (only GPU drivers) the problem lingered
Some fixes helped temporarily:
- Replacing RAM/PSU/CPU/Motherboard still didn’t resolve it
- Changing GPU to another model stopped Chrome crashes but not the freeze issue
- Replacing GPUs in other cases also failed to fix the core problem
Currently, there’s no clear solution, and it seems the system might need a full hardware overhaul or a complete platform replacement.
Probably haven’t tried every possible troubleshooting method. Feeling worn out to the point of giving up doesn’t mean all steps are done.
Inspect power cable connections, paying close attention to the 12VHPWR GPU cable for any signs of damage or poor contact.
Perform separate stress tests on the GPU and CPU using tools such as OCCT, FurMark, or Heaven Benchmark to ensure stability under load.
Assess the health of your storage devices with CrystalDiskInfo to exclude SSD/HDD problems that might cause delays leading to GPU resets.
Review NVIDIA TDR (Timeout Detection and Recovery) settings and consider increasing the TDR delay in the registry.
Make sure all motherboard chipset and storage drivers are up to date, not just reverting to an older BIOS version.
Attempt a clean Windows installation on an external drive to determine if the issue stems from OS corruption or configuration errors.
If feasible, test the GPU on another system to rule out a hardware defect.
To adjust and extend the NVIDIA TDR (Timeout Detection and Recovery) delay in Windows:
Open the Registry Editor by pressing Windows + R, typing regedit and pressing Enter.
Navigate to the TDR registry key under HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers.
Change or create the TdrDelay value in the right pane. If it exists, double-click it and set a higher decimal number (default is 2 seconds; a suggested setting is 60).
If it doesn’t exist, right-click in the right pane, select New –> DWORD (32-bit) Value, name it TdrDelay, double-click it, set the base to Decimal, and enter the value (e.g., 60).
(Optional) Adjust or create TdrDdiDelay for a longer allowance (default 5 seconds; can also be set to 60).
Follow these steps, then close the Registry Editor and restart your PC to apply changes.
This adjustment slows Windows from detecting an unresponsive GPU before it resets the driver, helping avoid freezes or black screens during intensive tasks.
Be careful: setting it too high may cause the system to hang longer if a genuine GPU failure occurs.
If problems continue, undo by deleting the keys or resetting values (TdrDelay=2, TdrDdiDelay=5).
What stress tests were performed?
Stress tested the PSU with HWiNFO for 15–20 minutes → everything normal.
Besides your patience? Did you run stress tests while using hwinfo and find no issues?
Welcome to the forums, newcomer!
Check out the updates for NVIDIA 572.83 (it’s been stable from March to July).
Have you tried any other driver updates besides 572.83?
DDU full reinstall of Nvidia drivers (I didn’t do this during my upgrade from 1080 Ti → 5080) plus I uninstalled Physx and blocked driver downloads from the Windows update section when searching for a device under "Windows". Then before starting DDU, I chose Clean and Shutdown. [20/08/2025]
My approach is to remove all GPU drivers (Intel, AMD, Nvidia) while in Safe mode, then install the newest GPU driver manually via an elevated command.
PSU:
Corsair HXi Series HX1000i 1000W
How long has your PSU been installed?
Crashes began:
July 2025
For clarity, what operating system are you using and which version (not edition) are you on?
Even after a clean Windows install (just GPU drivers) → the issue continued.
Troubleshooting tip:
every solution fits different people. What didn’t work for you may work for someone else. If your PSU seems to be the cause, try borrowing a well-built PSU and then restart the system. Also, think about reinstalling the OS after creating a bootable USB drive, but install the OS and drivers while offline.
How are you managing your processor cooling?
Did you receive your 13900k replaced or verify if you meet the criteria for a replacement because the 13k/14k series CPUs can fail over time, leading to crashes or bugs? Your problem differs from what others discussed about their 13k and 14k CPUs failing. They had to initiate a warranty RMA replacement with Intel. I'm not claiming the CPU is the sole cause, but it might be worth checking if you qualify for a warranty claim and replacement CPU.
I have only attempted one driver update that was installed directly on the PC without any additional steps (576.88 according to the latest catalog). I have heard positive updates about the 580.xx series and plan to test them as well.
The PSU in your build is from January 2023, but it's the 2022 model with ATX 2.52 specifications. It was purchased before Corsair updated it to the newer ATX 3.0 version.
You are using Windows 10 Pro, and if you're on Windows 11, please specify the OS edition you're running.
For cooling your processor, I'm using the NZXT Kraken X model.
I recently checked my PSU and discovered that even though I bought it in early 2023, it wasn't the refreshed ATX 3.0 version released later in 2023, but the older ATX 2.52 variant. I haven't installed it yet; I intend to do so later this month. It has passed all tests from the Intel Processor Diagnostic Tool 64, which was released to verify CPU health.