FWIW you can get better results using `Map()` in node. For example with node v8.3.0, using `Map` gives me 16.7 million items before the app crashes due to OOM.
Also for node, you should use the high-resolution timer (`process.hrtime()`) instead of `Date.now()` since the latter can be slow(er) due to calls to gettimeofday() IIRC.
Still kinda disappointing imho, how long did the test run before reaching this point? Is it also faster or kinda the same?
Also wouldn't a set be more appropriate in this case? Since the lookup time in a set would be faster than a map? Maybe a set also fits more ints?