PC Shuts Down While Gaming: Heat, Power or Something Else

A shutdown is not the same as a crash

If the game closes or the screen freezes, Windows is still involved and the fault is likely software, driver or memory related. If the machine goes dark instantly like the plug was pulled, Windows never got a say — something cut power. That distinction decides where to look.

• Goes dark and restarts by itself — often power instability or a protection trip.
• Goes dark and stays off until you press the button — commonly thermal protection or the power supply latching off.
• Goes dark only during specific heavy scenes or benchmarks — points at load spikes rather than a steady-state problem.

Thermal shutdowns

Processors and graphics cards throttle first and shut down as a last resort. A machine that reaches shutdown temperature normally shows warning signs beforehand: fans ramping to maximum, frame rates dropping after a few minutes, the case exhaust noticeably hot.

• Dust in heatsink fins and filters — the most common cause and the cheapest to fix.
• Thermal paste that has dried out, especially on machines several years old.
• A cooler fan or pump that has stopped or slowed.
• A cooler that has lost mounting pressure, sometimes after transport.
• Case airflow: intake blocked against a wall, filters clogged, or fans fitted the wrong way around.
• Ambient temperature — a machine that is marginal in winter can start shutting down in an Omaha-area summer.

Power supply and power delivery

Graphics cards draw in spikes far above their rated average. A supply that is old, low quality, or sized for a previous card can pass every idle test and still trip during those spikes. Overcurrent and overpower protection then shuts the system down deliberately.

• A supply several years old that has not been changed across two or three GPU upgrades.
• Daisy-chaining one cable's two connectors into a card that wants two separate cables.
• Mixing modular cables from a different power supply — pinouts are not standard between brands and this destroys hardware.
• A failing wall outlet, extension lead or overloaded surge strip, especially with a space heater or window unit on the same circuit.

What you can check safely

1. 1

   Log temperatures right up to the shutdown

   Use monitoring software that writes to a log file — an on-screen readout disappears with the machine. Look at the last recorded GPU and CPU temperatures.

2. 2

   Check Event Viewer

   A Kernel-Power event ID 41 confirms the machine lost power without shutting down cleanly. It confirms the type of failure; it does not name the part.

3. 3

   Clean the machine

   With it unplugged, clear dust from filters, radiator or heatsink fins and fan blades using compressed air. Hold fans still so they do not spin freely while you blow them.

4. 4

   Confirm every fan actually spins

   Watch during a load, including the pump on a liquid cooler. A stopped pump raises temperatures within seconds.

5. 5

   Move to a known-good outlet

   Bypass the surge strip and plug straight into the wall, on a circuit without heavy appliances.

6. 6

   Cap the load and see what changes

   Set a frame rate limit or lower settings. Stability at reduced load supports a thermal or power cause rather than a software one.

7. 7

   Return everything to stock

   Disable overclocks, undervolts and memory profiles. An unstable tune can present as a hard shutdown.

What each result may indicate

• Peak temperatures above 95 °C before shutdown — cooling is at least part of the problem.
• Temperatures in the seventies and still shutting down — heat is unlikely to be the cause; look at power delivery.
• Stable with a frame cap, shuts down uncapped — the machine fails at peak draw, which fits either the power supply or a card pulling more than it can be given.
• Shuts down under CPU load alone as well as gaming — points at the supply, the board or the outlet rather than the graphics card.
• Cleaning fixed it for a few weeks, then it returned — cooling is marginal rather than fixed; paste, cooler mounting or airflow needs proper attention.

When to stop DIY

Stop if the shutdowns start happening at idle or during startup, if you smell anything hot or electrical, or if the machine will not restart immediately afterwards. Those are signs of failing hardware rather than a protection trip, and continuing to power-cycle can make it worse.

Also stop when the remaining step is "buy a bigger power supply and see". That guess is right often enough to be tempting and wrong often enough to be expensive.

When to bring it in

On the bench we can load the machine while watching temperatures and substitute a known-good supply to settle the question directly. If it turns out to be thermal, cleaning, fresh paste and correcting airflow is straightforward work — and much cheaper than the parts you would otherwise buy on suspicion.