One Billion Row Challenge

My implementation of the One Billion Row Challenge in Rust.

Data Generation

TODO: Explain whats going on in data/src/main.rs

Benchmarks

TODO: Write script that benchmarks 1..=64 threads

TODO: Add graph comparing to other solutions

Dependencies

Yes this implementation has dependencies, which is technically not allowed ... but I am also not writing this in Java and the competition is over anyways.

I am using memmap2/libc to access mmap, madvise and memchr, which is fair imo because they just wrap standard functions.

rustc_hash ehhhh yeah, no excuses here really. Just didn't feel like handrolling a hash function.