/pyperf

Toolkit to run Python benchmarks

Primary LanguagePythonMIT LicenseMIT

pyperf

Latest release on the Python Cheeseshop (PyPI) Build status of pyperf on GitHub Actions

The Python pyperf module is a toolkit to write, run and analyze benchmarks.

Features

  • Simple API to run reliable benchmarks
  • Automatically calibrate a benchmark for a time budget.
  • Spawn multiple worker processes.
  • Compute the mean and standard deviation.
  • Detect if a benchmark result seems unstable.
  • JSON format to store benchmark results.
  • Support multiple units: seconds, bytes and integer.

Usage

To run a benchmark use the pyperf timeit command (result written into bench.json):

$ python3 -m pyperf timeit '[1,2]*1000' -o bench.json
.....................
Mean +- std dev: 4.22 us +- 0.08 us

Or write a benchmark script bench.py:

#!/usr/bin/env python3
import pyperf

runner = pyperf.Runner()
runner.timeit(name="sort a sorted list",
              stmt="sorted(s, key=f)",
              setup="f = lambda x: x; s = list(range(1000))")

See the API docs for full details on the timeit function and the Runner class. To run the script and dump the results into a file named bench.json:

$ python3 bench.py -o bench.json

To analyze benchmark results use the pyperf stats command:

$ python3 -m pyperf stats telco.json
Total duration: 29.2 sec
Start date: 2016-10-21 03:14:19
End date: 2016-10-21 03:14:53
Raw value minimum: 177 ms
Raw value maximum: 183 ms

Number of calibration run: 1
Number of run with values: 40
Total number of run: 41

Number of warmup per run: 1
Number of value per run: 3
Loop iterations per value: 8
Total number of values: 120

Minimum:         22.1 ms
Median +- MAD:   22.5 ms +- 0.1 ms
Mean +- std dev: 22.5 ms +- 0.2 ms
Maximum:         22.9 ms

  0th percentile: 22.1 ms (-2% of the mean) -- minimum
  5th percentile: 22.3 ms (-1% of the mean)
 25th percentile: 22.4 ms (-1% of the mean) -- Q1
 50th percentile: 22.5 ms (-0% of the mean) -- median
 75th percentile: 22.7 ms (+1% of the mean) -- Q3
 95th percentile: 22.9 ms (+2% of the mean)
100th percentile: 22.9 ms (+2% of the mean) -- maximum

Number of outlier (out of 22.0 ms..23.0 ms): 0

There's also:

  • pyperf compare_to command tests if a difference is significant. It supports comparison between multiple benchmark suites (made of multiple benchmarks)

    $ python3 -m pyperf compare_to --table mult_list_py36.json mult_list_py37.json mult_list_py38.json
    +----------------+----------------+-----------------------+-----------------------+
    | Benchmark      | mult_list_py36 | mult_list_py37        | mult_list_py38        |
    +================+================+=======================+=======================+
    | [1]*1000       | 2.13 us        | 2.09 us: 1.02x faster | not significant       |
    +----------------+----------------+-----------------------+-----------------------+
    | [1,2]*1000     | 3.70 us        | 5.28 us: 1.42x slower | 3.18 us: 1.16x faster |
    +----------------+----------------+-----------------------+-----------------------+
    | [1,2,3]*1000   | 4.61 us        | 6.05 us: 1.31x slower | 4.17 us: 1.11x faster |
    +----------------+----------------+-----------------------+-----------------------+
    | Geometric mean | (ref)          | 1.22x slower          | 1.09x faster          |
    +----------------+----------------+-----------------------+-----------------------+
    
  • pyperf system tune command to tune your system to run stable benchmarks.

  • Automatically collect metadata on the computer and the benchmark: use the pyperf metadata command to display them, or the pyperf collect_metadata command to manually collect them.

  • --track-memory and --tracemalloc options to track the memory usage of a benchmark.

Quick Links

Command to install pyperf on Python 3:

python3 -m pip install pyperf

pyperf requires Python 3.7 or newer.

Python 2.7 users can use pyperf 1.7.1 which is the last version compatible with Python 2.7.

pyperf is distributed under the MIT license.

The pyperf project is covered by the PSF Code of Conduct.