Say I have a very large data structure who’s size is in the 10s of MB. This structure is arbitrarily complex, consisting of many tuples and proplists and maps, but in a roughly tree-like structure. What is the most efficient way to compare two of these structures? Was curious before I started running some benchmarks. Is pattern matching breadth first, or depth first, or does it involve a bit to bit comparison of the terms? Would using zlib on a a term_to_binary then comparing the compressed binaries make sense?

I’m having trouble finding documentation on the generally most efficient way of doing this.