memory allocation of different js objects

ObjectCreationPerformance.js

window.__MemoryTest__ = {
  proto:[],
  concat:[],
  creat:[],
  newFn:[],
  factory:[]
};

function proto(){return new Foo()}
function concat(){return Object.assign({},baseObj)}
function creat(){return Object.create(baseObj)}
function newFn(){return ({fn:function(){}})}
function factory(){return ({fn:baseObj.fn})}

var creatorFns = {proto,concat,creat,newFn,factory};

var baseObj = {
  fn:function(){}
};

function Foo(){}
Foo.prototype = baseObj;

Object.keys(creatorFns).forEach((name)=>{
  const fn = creatorFns[name];
  const resultArray = window.__MemoryTest__[name];
  let loop = 100000;
  console.profile(name);
  while(loop--) resultArray[loop] = fn();
  console.profileEnd(name);
});
// prototypes mostly useful when objects are thrashing lots of memory
// like in 60fps games creating many objects... which few people ever do
// or tables with millions of rows
// even then, there are other ways to gain performance improvements,
// - referencing other objects
// - object pools
// - pass in the fns you need
// - optimize IO since web apps are usually IO bound
// arrayName, arrayBytes, oneObjBytes
// newFn    11316240    104 // ~9 million objects per gb
// creat     6515872     56 // ~18 million objects per gb
// concat   6515784     56 // ~18 million objects per gb
// factory  4115784     32 // ~31 million objects per gb
// proto    3315784     24 // ~42 million objects per gb

// Oddly, factory's objects are more space efficient than Object.create's, despite the prototype.
// also odd is the array size discrepancy with concat and creat
// factory seems to have the the best size performance if new objects are needed.
// no time to test execution performance tonight