Uploaded byMathias Bynens
Using jsPerf correctly
This document discusses best practices for using jsPerf, a tool for benchmarking JavaScript performance. It outlines common pitfalls to avoid like inaccurate timers, browser bugs, lack of statistical significance, unfair comparisons, asynchronous tests, and not properly setting up, preparing, and tearing down tests. The key recommendations are to only include the minimum code in each test, reset variables between tests, avoid randomness, don't test asynchronous code synchronously, and understand how Benchmark.js works for advanced usage. Following these guidelines will produce more accurate and reliable benchmark results.
![UiPath Automation Suite Installation (Hands-On) [2/3]](https://cdn.slidesharecdn.com/ss_thumbnails/automationsuitecommunitysession2-251015095633-a6d862f1-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
Using jsPerf correctly
- 1. Using jsPerf correctly @mathias– #jsconfeu 2011
- 2.
- 3.
- 4.
- 10. jsPerf is 426days old. 24,164 test cases → ~56.72 per day 83,536 tests → ~3.46 per test case 1,012 comments → ~2.38 per day
- 11. jsPerf is 426days old. 24,164 test cases → ~56.72 per day 83,536 tests → ~3.46 per test case 1,012 comments → ~2.38 per day
- 12. “68% of alljsPerf test cases is broken.” — Jean-Jacques Rousseau
- 13. “It would beunwise to take all jsPerf results for granted.” — William Shakespeare
- 14. Benchmarking pitfalls “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 15. Benchmarking pitfalls • Inaccuratemillisecond timers “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 16. Benchmarking pitfalls • Inaccuratemillisecond timers • Browser bugs “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 17. Benchmarking pitfalls • Inaccuratemillisecond timers • Browser bugs • No statistically significant results “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 18. Benchmarking pitfalls • Inaccurate millisecond timers • Browser bugs • No statistically significant results • Browser plugins and addons “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 19. Benchmarking pitfalls • Inaccurate millisecond timers • Browser bugs • No statistically significant results • Browser plugins and addons • No cross-browser testing “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 20. Benchmarking pitfalls • Inaccurate millisecond timers • Browser bugs • No statistically significant results • Browser plugins and addons • No cross-browser testing • Incorrect tests “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 21. Benchmarking pitfalls • Inaccurate millisecond timers • Browser bugs • No statistically significant results • Browser plugins and addons • No cross-browser testing • Incorrect tests “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 22. Benchmarking pitfalls • Inaccurate millisecond timers • Browser bugs • No statistically significant results • Browser plugins and addons • No cross-browser testing • Incorrect tests “Bulletproof JavaScript benchmarks” — http://mths.be/asv
- 23.
- 24. Function declarations go inthe preparation code or in the setup.
- 26. Only the bareminimum goes inside each test body.
- 28.
- 31.
- 33. Make sure themethods you’re comparing actually do the same thing.
- 34. If not, mentionthis in the description of the test case.
- 36. Make sure you’retesting what you intended to test.
- 37. http://jsperf.com/jquery-id-vs- native-getelementbyid
- 39. When re-using variables acrosstests, make sure you reset them as needed.
- 40.
- 43. Don’t introduce randomness inyour tests.
- 44.
- 45. Don’t test asynchronous stuffsynchronously.
- 46.
- 48. Asynchronous tests —a good example: http://jsperf.com/smallest-timeout
- 50. There’s a difference betweenpreparation, setup and teardown.
- 54. Example: scope lookups http://jsperf.com/truth
- 57.
- 61. jsPerf uses Benchmark.js, whichallows for advanced usage. http://benchmarkjs.com/docs
- 62.
- 64. This messes withyour results, too:
- 65. This messes withyour results, too: • Unexpected browser features
- 66. This messes withyour results, too: • Unexpected browser features • Dead code removal
- 67. This messes withyour results, too: • Unexpected browser features • Dead code removal •…
- 68. Questions? @mathias ~@jsprf