Start with the dump
What this fixes: PC crashes while gaming that dump to a blue screen, hard-reset the box, or kick you to desktop with a driver timeout. Who it’s for: anyone who would rather rule out drivers and heat before ordering a new GPU. Short version: capture the stop code, clean-install the GPU driver, then check thermals and power.
Microsoft publishes a bug-check code reference for reading stop codes after a blue screen. Write down the code (or photograph the screen) before you reboot past it — guessing without that string wastes hours.
Clean GPU drivers first
A large share of in-game crashes are stale or half-upgraded GPU drivers. AMD documents a factory-reset / clean-install FAQ (GPU-601) that removes the old stack before you install the current Adrenalin build. On Nvidia systems, download the current Game Ready driver from Nvidia’s GeForce drivers page and do a custom clean install rather than stacking an upgrade on a months-old package.
Driver releases themselves can reintroduce crash bugs — we have tracked that pattern before, including AMD’s Adrenalin 26.8.1 crash-fix notes where a single Adrenalin drop fixed three crash items and still left others open. If crashes started the day after a driver or Windows update, roll back that change before you replace hardware.
Windows patches can also take games down. our note on a Windows 11 patch crashing several games is one recent example of a security update colliding with live titles — so treat “it started this morning after Update Tuesday” as a real lead, not superstition.
Heat, power, and memory next
| Check | What to look for | Why |
|---|---|---|
| GPU / CPU temps | Thermal limit or sudden throttle | Heat trips protection mid-game |
| PSU headroom | Transient spikes on modern GPUs | Undersized units brown out under load |
| RAM stability | XMP/EXPO errors, mismatched kits | Unstable memory looks like random crashes |
| Event Viewer | Display / WHEA / Kernel-Power | Points at GPU, CPU, or PSU paths |
Intel’s support pages still treat unexpected shutdowns under load as a thermal and power-delivery problem first — see Intel’s unexpected-shutdown guidance for the processor-side framing. If clocks and fans look fine but Kernel-Power 41 events pile up, suspect the PSU or a wall/outlet issue before you RMA the GPU.
- Reseat power cables on the GPU and PSU side; pinched 12VHPWR/12V-2×6 adapters are a known failure mode on high-draw cards.
- Test with XMP/EXPO off for an evening. If crashes vanish, your memory kit or board firmware needs attention.
- Verify game files on Steam/Epic/Battle.net after a dirty driver install — corrupted shader caches mimic hardware faults.
When to stop and replace parts
If a clean driver, stock memory, sane thermals, and a known-good PSU still crash in every DirectX title, then hardware fault isolation (GPU in another PC, or a loaner card in yours) is the next step. GearChamber’s PC Builder helps you sanity-check whether the rest of the build is even rated for the card you are about to buy as a replacement.
Our take
We would not buy a new GPU until a clean driver install and a thermal/power pass say the old one is actually dying. Most “my PC crashes while gaming” threads that resolve do so on software and power — not a cart checkout.
Steps framed from AMD GPU clean-install FAQ, Nvidia GeForce drivers, Intel unexpected-shutdown guidance and Microsoft bug-check codes, checked 4 November 2026.
