Benchmark: Add item to array: push vs spread vs assign vs items.length vs items[len++]

Add item to array: push vs spread vs assign vs items.length vs items[len++] (version: 7)

Script Preparation code:

window.MAX_SIZE = 1000;

Tests:

Push item to an array
const items = []; for (let i = 0; i < MAX_SIZE; i++) { items.push(Math.random()); }
Append an item by spread
let items = []; for (let i = 0; i < MAX_SIZE; i++) { items = [...items, Math.random()]; }
Direct assign an item to an initialized array
const items = new Array(MAX_SIZE); for (let i = 0; i < MAX_SIZE; i++) { items[i] = Math.random();; }
items[items.length]
const items = []; for (let i = 0; i < MAX_SIZE; i++) { items[items.length] = Math.random(); }
items[items.length] outer length counter
const items = []; let len = 0; while (len < MAX_SIZE) { items[len++] = Math.random(); }
items[items.length] outer length counter + new Array()
const items = new Array(MAX_SIZE); let len = 0; while (len < MAX_SIZE) { items[len++] = Math.random(); }
Cutting empty tail
const items = new Array(1500); let len = 0; while (len < MAX_SIZE) { items[len++] = Math.random(); } items.length = MAX_SIZE;

Rendered benchmark preparation results:

Previous results

Test case name	Result
Push item to an array
Append an item by spread
Direct assign an item to an initialized array
items[items.length]
items[items.length] outer length counter
items[items.length] outer length counter + new Array()
Cutting empty tail

Latest run results:

Run details: (Test run date: one year ago)

User agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/68.0.3424.0 Safari/537.36

Browser/OS: Chrome 68 on Linux

Test name	Executions per second
Push item to an array	36145.0 Ops/sec
Append an item by spread	65.9 Ops/sec
Direct assign an item to an initialized array	74115.5 Ops/sec
items[items.length]	44867.3 Ops/sec
items[items.length] outer length counter	44740.8 Ops/sec
items[items.length] outer length counter + new Array()	71488.5 Ops/sec
Cutting empty tail	64240.8 Ops/sec

Related benchmarks: