Post #1834416

https://engineering.block.xyz/blog/18mb-of-hashmap-overhead The real question here is why in the world did someone decide to implement an array of 16 floats as a hashmap in the first place. The next question is, now that you've reduced it to a 64byte array instead of 792 bytes of hashmap and related goop, should be "how many of those 24674 arrays are identical?"... that's still 1.5MB (not counting heap overhead). ProTip: Think about the code you're writing. That's not evil "premature optimization", that's just common sense.