Skip to content

Advanced Topics

This section covers the internals of how NBenchmark works - for readers who want to understand the statistics behind the output, or who need to tune the library for unusual situations.

In this section

  • Statistics Deep Dive - how every number in the output is calculated, including the measurement loop, confidence intervals, and the significance test.
  • Validation & Accuracy - how those numbers are verified against SciPy/NumPy and to what tolerance.

Released under the MIT License.