An implementation of BitSet in Rust.
All is fully tested, including random tests with quickcheck.
I made simple benchmarks using cargo bench.
They check the speed of some simple bit operations on bitsets implemented as:
- a single u64 variable
- an array with a single u64 variable
- a list with a single u64 variable
There are two versions of each function checked: inlined and not inlined.
The full results are at report/Bits/report/index.html.
The interesting parts are:
- the speed of the not inlined vector is 4 times smaller than the speed of the rest of the functions
- the speed of the inlined vector implementation is the same as the speed of the single variable one