[bcrack]

The most appropriate compression algorithm to use is highly case-dependent (data structure, compression vs decompression cpu/wall-clock time, compression ratio, system memory,bandwidth, etc.). I'm sure people at kernel.org have their reasons to convert to xz.

The following link has some interesting (albeit non-exhaustive) benchmarks, comparing various compression algorithms (both serial and parallel implementations). http://vbtechsupport.com/1614/