The three approaches differ somewhat, with the binary search being the established state-of-the-art type algorithm. In fact, it can be used on any Rust slices (if they are sorted, of course) and used to find whatever is required.
Comparing these algorithms is tricky: a linear search works well on unordered datasets and is the only way to search those if sorting is not an option. If sorting is an option, then a binary search is faster by a large margin (asc is the sorting direction: ascending):
test tests::bench_binary_search_10k_asc ... bench: 80 ns/iter (+/- 32)
test tests::bench_binary_search_1k_asc ... bench: 63 ns/iter (+/- 17)
test tests::bench_binary_search_5k_asc ... bench: 86 ns/iter (+/- 28)
test tests::bench_jump_search_10k_asc ... bench: 707 ns/iter (+/- 160)
test tests::bench_jump_search_1k_asc ... bench: 92 ns/iter (+/- 10)
test tests::bench_jump_search_5k_asc ... bench: 355 ns/iter (+/- 46)
test tests::bench_linear_search_10k_asc ... bench: 2,046 ns/iter (+/- 352)
test...
