HWiNFO Sensor Log Analyser

How this reads your output

Temperature, clock and power columns are located by name, then summarised across the run. Logged throttling flags are counted where the sensor exists, and throttling is also inferred independently by looking for samples where the clock is well below its average while temperature is at its peak. That second check matters because many boards do not log a throttle flag at all, and a clock quietly held down is the effect you actually care about.

What the results mean

Peak against average

A brief peak during a benchmark is normal. A high average across a long run is what indicates cooling that is not keeping up.

Logged throttling events

Samples where the board reported thermal throttling. Absence does not mean it did not happen, since not every board exposes the sensor.

Samples clocked low while hot

Throttling inferred from behaviour rather than from a flag. This catches the case where no throttle sensor is logged.

Common problems and fixes

Temperatures hit 95 °C and I assume something is wrong

Modern processors are designed to run to their limit and reduce clocks briefly rather than overheat, so a peak of 95 °C under sustained load is normal behaviour rather than a fault. What matters is whether the clock is being held down for long periods, which is what the inferred throttling count above measures.

The log has no throttling column at all

Many boards do not expose one, which is why the clock behaviour is checked independently here. Look at the temperature plot alongside the average clock: a clock that falls and stays down as temperature reaches its peak is throttling regardless of whether anything flagged it.

Frequently asked questions

More in Interpreter

• CPU-Z / GPU-Z Report InterpreterRead a CPU-Z report and spot the two numbers everyone misreads.
• dxdiag Report InterpreterFind the four lines in a DxDiag report that explain why a game will not run.
• systeminfo / msinfo32 InterpreterTurn Windows systeminfo output into a readable spec sheet.