Unroll simple loop abstractions

[megawac]

Figure there is no reason to not write these simple loops our selves.

Latest tag is  v2.0.1
Comparing v2.0.1 with current on Node v6.2.2
--------------------------------------
queue(1000) v2.0.1 x 345 ops/sec ±0.80% (30 runs sampled), 2.90ms per run
queue(1000) current x 352 ops/sec ±0.90% (30 runs sampled), 2.84ms per run
current is faster
--------------------------------------
queue(30000) v2.0.1 x 9.39 ops/sec ±3.62% (18 runs sampled), 106ms per run
queue(30000) current x 10.52 ops/sec ±2.89% (20 runs sampled), 95.1ms per run
current is faster
--------------------------------------
queue(100000) v2.0.1 x 2.53 ops/sec ±11.46% (6 runs sampled), 396ms per run
queue(100000) current x 3.05 ops/sec ±8.64% (6 runs sampled), 328ms per run
current is faster
--------------------------------------
queue(200000) v2.0.1 x 1.26 ops/sec ±16.98% (3 runs sampled), 797ms per run
queue(200000) current x 1.43 ops/sec ±11.58% (3 runs sampled), 698ms per run
current is faster
--------------------------------------
current faster overall (1120ms total vs. 1300ms total)
current won more benchmarks (4 vs. 0)