Implementation benchmark v1.3.1.0

These are the compiled microbenchmark performance files, which compare multiple different implementations of the same operation. Each of them runs the same non-inlined implementation many times using a pregenerated ~8k element array of random values (wrapping to the start when needed). I used them to determine which version to use in this library based on their performance across a range of microarchitectures.

Windows: GCC 14.2, Linux: GCC 12.1, both using -O2.