Is your Ryzen actually overheating?
Ryzen chips are designed to boost until they reach their temperature limit. High numbers under heavy load are not automatically a fault. Real overheating usually shows up as:
- Clock speeds falling under sustained load, which is thermal throttling
- Sudden shutdowns or restarts
- Unusually high temperatures at idle, not just under load
Read the CPU (Tctl/Tdie) sensor in HWiNFO64 under a real load, and compare the number with the limits in the table below. Then check the clock speed, because a high number with steady clocks is usually normal behaviour.
Why Ryzen runs hotter than you expect
- The boost limit is the target. Ryzen 7000 and 9000 chips boost until they reach 95°C under all-core load. AMD describes this as normal, and the chip uses the headroom to boost.
- Idle spikes are common. Small, dense chiplets can produce quick spikes of 10–15°C at idle. A brief spike is boost behaviour, not a bad mount. A fan curve with a short step-up delay, such as 3–5 seconds, stops the fan from revving on every spike.
- Bigger coolers may not change the peak much. The chip can use extra cooling headroom to boost further. The benefit often shows up as quieter fans or steadier boost clocks rather than a much lower peak.
- AM5 chips run a little warmer. Their heat spreader is thicker, so AM4 coolers still fit.
Ryzen temperature limits by family
These are the maximum operating temperatures (TjMax) that AMD lists. Check AMD’s product page for your exact model.
| Chip family | TjMax (throttle point) |
|---|---|
| Ryzen 5000 | 90°C |
| Ryzen 7000 and 9000, non-X3D | 95°C |
| Ryzen 7800X3D and 7950X3D | 89°C |
| Ryzen 9800X3D | 95°C |
For what the limit means, see what TjMax means.
Fix 1: Cooler mounting, paste and dust
Start here, because these fixes cost nothing and address the most common causes.
- Remove the protective film. If your cooler has a plastic film on the cold plate, make sure it is off.
- Check the mounting pressure. The cooler should sit flat with even, firm pressure. Tighten the mounting screws diagonally in small steps.
- Repaste. Use a pea-sized dot or an X pattern. More paste is not better. On AM5, excess paste can squeeze into the notches around the heat spreader.
- Clean the dust. Blow dust out of the heatsink fins, the fan filter and any radiator. Hold the fans still while you use compressed air so they do not spin up.
- Check an AIO liquid cooler. The pump should run at full speed on its header. The pump should not be the highest point of the loop, because air gathers there.
Liquid metal is not needed for most Ryzen chips. The gain is small and the spill risk is real. Never use it near the socket, because it conducts electricity.
Fix 2: Eco Mode or a PBO thermal limit
- Eco Mode lowers the chip’s power target to 105 W or 65 W. It can cut temperatures a lot, with a small loss in multi-core performance. You can set it in AMD’s Ryzen Master software or in BIOS.
- PBO Thermal Limit in BIOS caps the temperature directly. Values around 80–85°C are common. It trades some peak performance for cooler running.
Fix 3: Curve Optimizer
Curve Optimizer applies a negative voltage offset to each core. Offsets around -10 to -30 are common, but stability varies from chip to chip. Change one value at a time, and test each step with a stress test and a few real games. If the system crashes, back off the offset. A lower voltage often means a lower temperature, but only if the chip stays stable.
Fix 4: Check the BIOS for the AM5 SoC voltage fix
In 2023, some Ryzen 7000 owners on AM5 boards reported bulging socket pads and burn marks. Some suspected EXPO memory profiles, which AMD had not approved for the 7000X3D chips at the time. AMD traced the problem to SoC voltage and released updated AGESA code that caps SoC voltage at about 1.3 V on affected boards. AMD said the change does not affect EXPO or XMP memory overclocking, or PBO boost.
Update the BIOS from your motherboard maker, then check that the setting is at its default if you have changed it. If you think a CPU was damaged, contact AMD customer support.
Fix 5: Fan curves and airflow
- Set a fan curve. Make sure the CPU fan is not stuck at a low fixed speed. The BIOS fan page or a fan control tool can set a curve that ramps up with temperature.
- Balance the airflow. Front and bottom intakes with rear and top exhaust help. A slight positive pressure, with more filtered air coming in than going out, reduces dust build-up.
- Keep the room cool. The room temperature adds roughly one-for-one to the CPU temperature. See PCs and room conditions for more.
When it is a real problem
Consider a real overheating problem when you see:
- Throttling under light or moderate load, not just in the longest stress tests
- Sustained high temperatures at idle, even after you have cleaned the cooler and checked the mounting
- Repeated shutdowns during play
Reset the BIOS to defaults, disable any aggressive overclocking, and run the fixes above in order. If the problem continues, check whether the CPU itself is damaged. See how to tell if a CPU is damaged from overheating.