
Start Here: Triage by Crash Type
Game crashes come in several flavors, and the diagnostic path is different for each:
- Crash to desktop with no error — usually drivers, overclocks, or VRAM
- Crash with “DXGI_ERROR_DEVICE_REMOVED” — GPU instability, often PSU or overclock
- Blue screen of death — hardware/driver issue, see the BSOD code
- System restart with no warning — almost always PSU or thermal
- System freeze (no input) — RAM or storage issue
- Black screen with continuing audio — GPU driver crash, often recoverable
Step 1: Reliability Monitor
Press Windows key, type “Reliability Monitor”, open it. This shows the past few weeks of crashes with timestamps. Patterns matter — crashes only during gaming point to GPU/PSU; crashes during file operations point to storage; crashes during sleep/wake point to drivers.
Step 2: Check Event Viewer
Press Win+X, click Event Viewer. Navigate to: Windows Logs → System. Filter by “Critical” and “Error” levels. Look for the timestamp of your most recent crash. Common culprits:
- Kernel-Power 41 — system rebooted without proper shutdown (PSU or thermal)
- WHEA-Logger errors — hardware fault (often CPU or motherboard)
- nvlddmkm or AMD display — GPU driver crash
- Application Error — usually a software issue specific to that game
Step 3: Update Drivers (the Right Way)
“Did you update your drivers?” is the most overused troubleshooting advice — but it’s first because it actually fixes a lot of issues.
- GPU: use DDU (Display Driver Uninstaller) in Safe Mode to fully remove old drivers, then install the latest stable (not beta) from NVIDIA, AMD, or Intel directly.
- Chipset: download from your motherboard manufacturer’s site, not Windows Update.
- Storage: check NVMe firmware on the SSD manufacturer’s site. Outdated firmware causes random Windows freezes.
Step 4: Test the PSU
This is the single most common cause of “random” gaming crashes. The PSU is barely making spec, and modern GPUs spike well above their rated power draw during transients.
How to test:
- Borrow a known-good PSU from a friend’s PC — the easiest test
- Disable Multi Frame Generation and any overclocks; if crashes stop, the PSU is the bottleneck
- Use HWInfo64 to monitor +12V rail under load — drops below 11.6V indicate PSU weakness
Step 5: Test the RAM
Memory errors cause random freezes, BSODs, and “data integrity” issues. Even XMP/EXPO profiles can be unstable.
- Run MemTest86 (free, bootable) for at least 4 hours. Any errors at all = bad stick or bad slot
- If you have XMP/EXPO enabled, try disabling it and running at JEDEC speeds. Many “random” crashes are actually unstable memory overclocks the manufacturer optimistically labeled stable
- If MemTest passes but games still crash, try Karhu RAM Test (paid) — it catches more subtle errors
Step 6: Test the CPU
Run Prime95 Small FFTs for 30 minutes. Watch temperatures with HWInfo. If the CPU hits 100°C and throttles, your cooler is undersized or your paste is dried out. Crashes during Prime95 with reasonable temps indicate CPU instability — often from Precision Boost Overdrive or other “auto-overclocking” features.
Step 7: Test the GPU
Run 3DMark Time Spy stress test (looped). Watch temps. Look for visual artifacts. A GPU that fails this test will crash games. A GPU that passes this test but still crashes specific games is usually a driver or game-specific issue.
Step 8: Storage Health
CrystalDiskInfo (free) reads SMART data from your drives. Look for “Caution” or “Bad” health status. Also check for unusual reallocated sector counts. A failing SSD or HDD causes games to hang, crash, or refuse to load assets.
Step 9: Storage Trim and Defrag
Windows usually handles this automatically, but verify. Open Defragment and Optimize Drives. Make sure your SSDs show as “OK” with recent optimization timestamps.
Step 10: Game-Specific Issues
If the crash is in only one game, the bug is in the game, not your PC. Common fixes:
- Verify game files in your launcher (Steam, Epic, etc.)
- Disable shader cache and clear it: ProgramData/NVIDIA Corporation/NV_Cache or similar
- Disable game overlays (Discord, GeForce Experience, Steam, Xbox Game Bar)
- Set the game’s executable to run in compatibility mode or windowed full-screen
- Check community forums for known crash issues — there are often game-specific fixes
Common Patterns and Their Causes
| Pattern | Likely Cause |
|---|---|
| Crash after 20-30 min of play | Thermal throttling or PSU |
| Crash during loading screens | Storage or RAM |
| Crash in specific game zones | VRAM exhaustion (lower texture settings) |
| Crash with audio looping | Display driver crash (update drivers) |
| Random BSOD across multiple uses | Hardware fault (RAM, CPU, or SSD) |
| Crash only when streaming/recording | Encoder overload or storage write speed |
The “Stress Test Everything” Day
If after the above checks you still have random crashes, set aside a Saturday to run:
- MemTest86 — 4 hours, no errors allowed
- Prime95 Small FFTs — 1 hour, no errors, no thermal throttling
- 3DMark Time Spy Stress Test — 20 loops, passing grade
- CrystalDiskMark on all drives — sustained reads/writes match spec
- HWInfo logged through all of the above — voltages stable, temps reasonable
If any of these fails, you’ve found your culprit. If all pass and games still crash, the problem is software/driver, not hardware.
When to RMA
If you’ve isolated the bad component, claim warranty. Most CPU and GPU manufacturers offer 2-3 year warranties. Motherboards and RAM are typically lifetime (within reason). Document the failure with screenshots, MemTest86 results, or HWInfo logs — RMA support is more responsive when you can show specific evidence.
Bottom Line
Random crashes are 90% PSU, RAM, drivers, or thermal. Work through this list in order and you’ll find the cause. The diagnostic process is the same whether your PC is a $700 budget build or a $5,000 enthusiast tower — components fail, software has bugs, and methodical elimination beats random Reddit suggestions every time.



