Add benchmarks about iterum

[xgbuils]

Hi @benji6,

Sorry, but the first test that I want to add is about slice method and I want to do some test with ramda. However I don't have so much experience with ramda and if it is possible to apply slice over a set iterable:

const set = new Set([1, 2, 3])
Array.from(iterum.slice(set, 0, 3)) // [1, 2, 3]
Array.from(imlazy.slice(0, 3, set)) // [1, 2, 3]
// and ramda?

I tried Array.from(ramda.slice(0, 3, set)) but the performance of ramda was really good and I discovered that the reason is that returns empty array [].

Maybe it is not applicable for ramda.

Thanks!

[xgbuils]

Hi!

Waiting for if it exists any way to slice a Set instance with ramda, I created some performance tests for slice.

For iterum 1.3.0 I have these results:

- slice
    - array - small start index
        - iterum x 102,306 ops/sec ±0.24% (90 runs sampled)
        - imlazy x 115,186 ops/sec ±5.50% (71 runs sampled)
        - ramda x 1,559,411 ops/sec ±0.27% (94 runs sampled)
        - native x 1,908,464 ops/sec ±0.91% (89 runs sampled)
        Fastest is native
    - array - big start index
        - iterum x 5,129 ops/sec ±0.71% (94 runs sampled)
        - imlazy x 24,571 ops/sec ±2.85% (84 runs sampled)
        - ramda x 1,510,341 ops/sec ±0.15% (89 runs sampled)
        - native x 1,878,096 ops/sec ±0.20% (93 runs sampled)
        Fastest is native
    - array - create new sliced iterable (no traversing)
        - iterum x 187,848 ops/sec ±1.60% (87 runs sampled)
        - imlazy x 3,280 ops/sec ±0.88% (94 runs sampled)
        - ramda x 1,013 ops/sec ±1.09% (70 runs sampled)
        - native x 1,028 ops/sec ±0.82% (71 runs sampled)
        Fastest is iterum
    - no indexable iterable - small start index
        - iterum x 85,268 ops/sec ±0.25% (89 runs sampled)
        - imlazy x 101,178 ops/sec ±4.33% (76 runs sampled)
        Fastest is imlazy
    - no indexable iterable - big start index
        - iterum x 3,723 ops/sec ±4.44% (92 runs sampled)
        - imlazy x 9,381 ops/sec ±1.93% (88 runs sampled)
        Fastest is imlazy
    - no indexable iterable - create new sliced iterable (no traversing)
        - iterum x 183,791 ops/sec ±1.26% (94 runs sampled)
        - imlazy x 1,079 ops/sec ±0.13% (92 runs sampled)
        Fastest is iterum

For iterum branch dev-2.0.0

- slice
    - array - small start index
        - iterum x 229,301 ops/sec ±1.16% (95 runs sampled)
        - imlazy x 117,158 ops/sec ±5.18% (75 runs sampled)
        - ramda x 1,568,007 ops/sec ±0.28% (96 runs sampled)
        - native x 1,941,122 ops/sec ±0.21% (93 runs sampled)
        Fastest is native
    - array - big start index
        - iterum x 221,244 ops/sec ±1.87% (90 runs sampled)
        - imlazy x 25,342 ops/sec ±2.24% (87 runs sampled)
        - ramda x 1,470,998 ops/sec ±0.28% (94 runs sampled)
        - native x 1,899,217 ops/sec ±1.54% (92 runs sampled)
        Fastest is native
    - array - create new sliced iterable (no traversing)
        - iterum x 274,096 ops/sec ±1.31% (88 runs sampled)
        - imlazy x 3,291 ops/sec ±0.73% (95 runs sampled)
        - ramda x 1,016 ops/sec ±1.22% (72 runs sampled)
        - native x 1,011 ops/sec ±0.98% (77 runs sampled)
        Fastest is iterum
    - no indexable iterable - small start index
        - iterum x 139,654 ops/sec ±1.45% (93 runs sampled)
        - imlazy x 102,659 ops/sec ±4.39% (76 runs sampled)
        Fastest is iterum
    - no indexable iterable - big start index
        - iterum x 8,083 ops/sec ±0.21% (90 runs sampled)
        - imlazy x 9,258 ops/sec ±2.10% (89 runs sampled)
        Fastest is imlazy
    - no indexable iterable - create new sliced iterable (no traversing)
        - iterum x 239,569 ops/sec ±1.44% (93 runs sampled)
        - imlazy x 1,055 ops/sec ±0.47% (90 runs sampled)
        Fastest is iterum
    - two calls
        - iterum x 7,968 ops/sec ±0.13% (95 runs sampled)
        - imlazy x 4,644 ops/sec ±1.86% (90 runs sampled)
        Fastest is iterum

[benji6]

Hey this is awesome!

So as far as I'm aware Ramda can't slice a Set... You could spread it into an array then pass it into Ramda's slice function first I suppose, but it's not quite the same thing

[xgbuils]

perf/benchmarks.md has some benchmarks. I have to work hard improving the performance after publishing release 2.0.