alt Hacker News> a small-multiple boost
Quick reminder that a 20x boost is better than going from O(n) to O(log n) for up to a million items. And, that log n algorithms often are simply not possible for many problems.
> a small-multiple boost
Quick reminder that a 20x boost is better than going from O(n) to O(log n) for up to a million items. And, that log n algorithms often are simply not possible for many problems.
Am I getting the math wrong here? Going from O(n) to O(log n) (with no change in constant factor) for a million items would be going from ~1000000c ops to ~20c ops, which would be a 50000x improvement?