I just published CRC32C routines and benchmarks that have been highly optimized for Edison.
The fastest is IntelC (a port to C of the 64bit hand optimized asm linux kernel routine) runs about 10x faster than just simple HW crc32l instructions would and 38x times faster than the fastest SW based algorithm (aka SlicingBy8).
The throughput for a 4k buffer reaches almost 2GiB/sec.
I hope this might be useful in your IIoT project.
On a 64bit Edison kernel this might be even 2x faster. Any ideas for further optimization are welcome.