Error message indicates a persistent issue, preventing gameplay beyond 30 minutes.
Error message indicates a persistent issue, preventing gameplay beyond 30 minutes.
I’m experiencing repeated crashes in most games after just 5 to 30 minutes of playtime, and I’m unsure what steps to take next. I was banned from Overwatch 2 and Marvel Rivals because I can’t complete any match without freezing. My setup includes:
- Ryzen 7 8700G (currently replaced with a 7700)
- Gigabyte Aorus B650 Elite AX ICE
- Asus RTX 4070 Ti Super OC
- Kingston 32 GB DDR5 CL40 6000
- SSD XTD 2TB
- Corsair 850W PSU (80+Gold)
When crashes occur, Windows Event Viewer displays the error code “nvlddmkm 153.” I suspect this driver issue might be linked to hardware problems. I’ve tried several fixes without success:
- Updated drivers and reverted to previous versions
- Used DDU for diagnostics
- Updated motherboard BIOS
- Changed TDR registry settings
- Performed a fresh Windows install
- Replaced the GPU (Asus) after support confirmed it works
- Ran tests on RAM, CPU, GPU, SSD using known tools
- Checked temperatures—all within normal ranges
- Disassembled and reassembled PC; old GPU worked fine
- Switched to a different GPU (Radeon RX 5700) eliminated crashes
- Reduced TDP on the 4070 but still experienced issues
- Noted that Overwatch and Marvel Rivals freeze, though audio plays briefly before timeouts
I tested every component individually, including RAM with Memtest86, CPU with Prime95, GPU with Furmark, and SSD with Crystaldisk. The RAM issue persists, so it’s unlikely to be a single component problem.
Recently, I considered the possibility that the combination of CPU, GPU, and RAM is causing instability. Since the 8700G isn’t a high-end processor, I thought upgrading to a more common model like the 7700 might help. However, even after changing CPUs, the crashes continue.
The error code suggests a communication failure between CPU and GPU, which aligns with my suspicion. It’s also possible that transient voltage spikes during gaming stress could damage the PSU briefly, triggering this error. I’ve read that synthetic tests don’t fully mimic real-world gaming loads, so they might not capture the exact issue.
Before purchasing a new PSU, consider:
- Verifying the PSU’s wattage matches your system’s needs
- Ensuring proper grounding and connections
- Checking for any signs of physical damage or loose components
- Running stress tests with tools like Prime95 or Furmark again to confirm stability
I’m still trying to figure out exactly what’s causing this, but it seems hardware interaction is likely. Patience and thorough testing are key here.
I agree with your assessment. Most of the other problems seemed linked to power supply issues. The faulty PSUs caused the PCIe cable to melt, leading to crashes and errors. The EVGA unit didn’t fail completely, but after replacing several components, it worked again. You might find a replacement cable on sites like eBay or Amazon, or consider higher-end options from brands like CableMod. If you know someone with a compatible PSU, ask them if they’d be willing to lend it for testing before proceeding with a return.
I've noticed that PSU issues can sometimes appear normal until they cause problems. I'm performing the OCCT power test, though I'm unsure if it accurately detects failures. Thus far, no issues have been found, but I plan to run it for an hour to confirm.
This has been confusing me. I used Furmark for more than 30 minutes without any issues—no crashes, no artifacts, no reboots, and the GPU was consuming significant power. I'm running the OCCT 3D Adaptive test with the switch mode enabled to simulate transients, and everything is proceeding smoothly so far.
Mentioned the problem clearly—Furmark would trigger a black screen in roughly five seconds when it started. Also, HWInfo was running in the background, which helped verify that my loads and temperatures stayed normal during the incident.