Crappy Node Benchmarks

This is a personal repo I use to do simplistic benchmarks on competing node libraries when performance is a criteria.

Pull requests for modifications and additional libraries are welcome, but the testing criteria must be the exact same as the competing libraries.

Why Crappy? Because it's simplistic and to cover the a common use case, not all use cases. I also use the word crappy liberally, I don't mean to offend the library authors, and every single one should be given credit for the effort they spend on these libraries.

Results

Category	Candidate	Test No.	Iterations	Sec	Millis
CLI	Yargs	1	1000	123 sec	123680 millis
CLI	Commander	1	1000	46 sec	46648 millis
CLI	argparse	1	1000	57 sec	57982 millis
Config	Convict	1	1000	89 sec	89192 millis
Config	Convict	2	1000	88 sec	88116 millis
Config	Nconf	1	1000	109 sec	109839 millis
Config	Nconf	2	1000	72 sec	72314 millis

CLI parsers

1000 iterations of <candidate>-bench add -s 10 1 3 5 7 11:

• Yargs: 123 sec, 123680 millis
• Commander: 46 sec, 46648 millis
• argparse: 57 sec, 57982 millis

More details

Run docker-compose -f docker-dir/cli-parsers-compose.yml

Config parsers

1000 iterations of <candidate>-bench<testNumber> -c 19 -d 23:

• Convict:
  • Test 1: Multiple files, with env and CLI override: 89 sec, 89192 millis
  • Test 2: Multiple files, no env or CLI override: 88 sec, 88116 millis
• Nconf:
  • Test 1: Multiple files, with env and CLI override: 109 sec, 109839 millis
  • Test 2: Multiple files, no env or CLI override: 72 sec, 72314 millis

More details

Run docker-compose -f docker-dir/config-parsers-compose.yml

Improvements

• Run on multiple Node versions, not just Node 4. It's unfair to criticize a libraries performance based on how it operates on an old version of a language, when it could very well leverage new functionality in newer versions.
• add run commands for non docker-compose.