Based on Stephan Brumme's Fast CRC32 library.

But with modified tables from Intel‘s Slicing-By-8, to match SSE4.2’s CRC32 instruction, which uses CRC-32C (also known as Castagnoli CRC32).