Skip to content

BENCHMARK METHOD

Do not compare a number without its measurement conditions.

This page does not state current performance or superiority. It defines the information needed for a reproducible comparison.

GUIDE

What to fix before measuring

Identify the targetRecord the media type, capacity, connection, condition, and the information used to identify it.
Control the environmentKeep the hardware, OS, tool version, power, connection, and concurrency conditions that can affect the result.
Define start and finishState whether preparation, recognition, execution, verification, and record transfer are inside the measured interval.
Separate exceptionsKeep retries, failures, exclusions, and manual intervention separate from summary values and record the reason.
Preserve results and evidenceKeep per-target records and the measurement date in addition to aggregate values.Review the evidence model

Review reproducibility before the number

When current measurements are required, fix the target and environment and test with the tool version in use.

How to measure erasure time | MASAMUNE AI