Skip to content

Latest commit

 

History

History
35 lines (27 loc) · 1.33 KB

BENCHMARKS.md

File metadata and controls

35 lines (27 loc) · 1.33 KB
Raw

Benchmarks

broccoli-concat includes some easy to run benchmarks for verifying the performance characteristics of the plugin. By default, you can simply do:

yarn run bench
# or
npm run bench

This will run benchmarks for initial builds and rebuilds. You can optionally provide several environmental variables to customize the benchmarking runs:

  • FILE_COUNT (number, default: 10) - Specifies how many files to concat together in the fixture.
  • DEPTH (number, default: 1) - Specifies how many sub-directories of files to place in the fixture. Each directory level will have FILE_COUNT files.
  • LARGE_FILES (boolean, default: false) - Specifies whether to use fixture files with large memory footprints (>1mb).
  • SOURCE_MAPS (boolean, default: true) - Specifies whether to use sourcemap concat.

So, if you want to run many iterations testing the performance of concatenating several large files in a nested structure, you'd run something like:

FILE_COUNT=3 DEPTH=4 LARGE_FILES=true yarn run bench

The resultant output should be something like:

Benchmarking with the following parameters
------------------------------------------
FILE_COUNT: 3
DEPTH: 4
LARGE_FILES: true
------------------------------------------
Average time for build over 10 iterations: 213.6ms
Average time for rebuild over 10 iterations: 288ms