Confirm whether clocks are being constrained
CPU thermal throttling is a reduction in operating frequency or power caused by a temperature limit. A high temperature alone is not enough; the diagnosis requires falling effective clocks, a limit indicator or reduced performance during a sustained repeatable workload.
A power, thermal or electrical limit is supported by falling effective clocks and a matching limit signal, not temperature alone. First hypothesis to test: Cooler contact can be poor.
Compare cold and sustained behavior
Use the first two controls in order: Record a cold baseline and warmed-up run; then Log effective clocks, package power and limits. This establishes whether the first observation survives a controlled comparison.
Run the checks in this order
- Record a cold baseline and warmed-up run.
- Log effective clocks, package power and limits.
- Repeat with safe default firmware settings.
- Inspect cooler mounting and airflow.
A cooling or power-policy change should restore the expected clock or completion behavior in the same workload.
Pair limit signals with effective clocks
Result, control and meaning
| Observation | Controlled change | Likely interpretation |
|---|---|---|
| Temperature rises, clocks fall | Correct cooling | Thermal throttling is supported |
| Clocks stable at high temperature | Check target performance | The CPU may be operating normally |
| Power limit triggers first | Review board settings | The constraint is not purely thermal |
When the observation is “Temperature rises, clocks fall,” apply “correct cooling.” The evidence table reads the response as follows: Thermal throttling is supported.
Heat, power policy and electrical limits to separate
Compare two early explanations: (1) Cooler contact can be poor; and (2) Dust and fan curves restrict heat removal. They can resemble one another in a headline result but predict different responses to the checks below.
Alternative causes to inspect
- Cooler contact can be poor.
- Dust and fan curves restrict heat removal.
- Motherboard power settings raise heat.
- Ambient temperature changes headroom.
Compare the cold result with a warmed, sustained pass while logging clocks, temperature, power and limit reasons together.
Restore normal operation before replacing parts
Restore expected operation before considering a faster processor. A new CPU placed under the same inadequate cooler can reproduce the problem. After maintenance, repeat the original workload and save the stabilized clock and temperature baseline.
Test record
Record effective clocks, package or board power, hotspot and core temperatures, fan behavior and every active limit flag.
- package temperature
- effective frequency
- thermal flag
- cooler contact
- fan curve
- cold pass
- soak period
- ambient reading
Checks that can overturn the decision
- Calling any high temperature throttling.
- Using requested instead of effective clocks.
- Testing only a short burst.
Decision boundary: The decision above should also survive the most likely shortcut: calling any high temperature throttling.
Technical sources
References for the measurement method
