priority queue needs optimizations / benchmarking

[kasima]

This has only been tested to few exits and gas costs go up as exits go up. Need to benchmark, make sure it works, and optimize, if needed.

Primarily concerned with insertions

[smartcontracts]

So we can achieve O(1) insert if we implement something like a Fibonacci heap. We're still basically limited by log(n) delete-min during the finalization of exits, but that's less of a concern than having a large gas cost be a limiting factor to the submission of exits.

[smartcontracts]

Closing as out of scope for MVP.