benchmarking with asv

[aaronspring]

• requires Refactor get keepbits #46
• required for bitround with xr.apply_ufunc #47 and bitround inplace  #42
• label run-benchmarks and GHA runs all benchmarks which are not marked slow
• checks timing and peak memory of functions: get_bitinformation, get_keepbits and xr/jl_bitround
• add contribution.rst (copied from climpred) with instructions for asv, PRs & pytest
• xr_bitround on rasm fails until Use numcodecs bitround #43 is merged
• julia calls fail on GHA, opened ReadOnlyMemoryError for julia/pycall when using asv on GHA #50

[aaronspring]

didnt see we already had a contibuting.rst, copied it over. which version do you like more?

[aaronspring]

fails the float64 tests, should run after #43 is merged. I would like to merge @observingClouds