Run the tests in parallel

[chasenlehara]

This improves the time it takes to run the exercise solution tests by running them in parallel.

Below are the results of multiple test runs in GitHub and on a MacBook Pro. It looks like about three tests running in parallel is optimal in both environments, so the test script will max out at running three tests at a time.

First run in GitHub CI:

┌─────────┬───────┬───────────────┬─────────────┐
│ (index) │ limit │ time          │ formatted   │
├─────────┼───────┼───────────────┼─────────────┤
│ 0       │ 4     │ 787608.631378 │ '13:07.608' │
│ 1       │ 3     │ 461751.497619 │ '7:41.751'  │
│ 2       │ 2     │ 475695.005603 │ '7:55.695'  │
│ 3       │ 1     │ 646006.242796 │ '10:46.006' │
└─────────┴───────┴───────────────┴─────────────┘

Second run (after changing the max limit after realizing the first run will be slower while priming the npm cache):

┌─────────┬───────┬───────────────┬─────────────┐
│ (index) │ limit │ time          │ formatted   │
├─────────┼───────┼───────────────┼─────────────┤
│ 0       │ 5     │ 763990.330012 │ '12:43.990' │
│ 1       │ 4     │ 432908.429822 │ '7:12.908'  │
│ 2       │ 3     │ 445064.747047 │ '7:25.064'  │
│ 3       │ 2     │ 468414.323292 │ '7:48.414'  │
│ 4       │ 1     │ 639965.18994  │ '10:39.965' │
└─────────┴───────┴───────────────┴─────────────┘

First run on a MacBook Pro:

┌─────────┬───────┬───────────────┬─────────────┐
│ (index) │ limit │ time          │ formatted   │
├─────────┼───────┼───────────────┼─────────────┤
│ 0       │ 8     │ 666539.938417 │ '11:06.539' │
│ 1       │ 7     │ 607849.072083 │ '10:07.849' │
│ 2       │ 6     │ 571706.310166 │ '9:31.706'  │
│ 3       │ 5     │ 516670.640167 │ '8:36.670'  │
│ 4       │ 4     │ 498700.136167 │ '8:18.700'  │
│ 5       │ 3     │ 471225.195584 │ '7:51.225'  │
│ 6       │ 2     │ 485589.121083 │ '8:05.589'  │
│ 7       │ 1     │ 617190.439583 │ '10:17.190' │
└─────────┴───────┴───────────────┴─────────────┘

Second run:

┌─────────┬───────┬───────────────┬─────────────┐
│ (index) │ limit │ time          │ formatted   │
├─────────┼───────┼───────────────┼─────────────┤
│ 0       │ 8     │ 653896.0615   │ '10:53.896' │
│ 1       │ 7     │ 578432.966958 │ '9:38.432'  │
│ 2       │ 6     │ 550287.004208 │ '9:10.287'  │
│ 3       │ 5     │ 509471.244167 │ '8:29.471'  │
│ 4       │ 4     │ 497503.24975  │ '8:17.503'  │
│ 5       │ 3     │ 468895.376833 │ '7:48.895'  │
│ 6       │ 2     │ 492863.924542 │ '8:12.863'  │
│ 7       │ 1     │ 615069.0665   │ '10:15.069' │
└─────────┴───────┴───────────────┴─────────────┘

[DanDaManTran]

Test is failing when running benchmark. It pass successfully without benchmark

[chasenlehara]

Test is failing when running benchmark. It pass successfully without benchmark

Interesting, I ran into this too at parallel limits 9+ (my Mac as 10 cores). I don’t think this matters for merging it in because I left the benchmark code mainly as a reference for doing something similar in the future.