I have a large compress task (compress 10000 large file( > 10 G each))
I found xz -z
or gzip
are slow.
So, i want to use multiprocess to make the compress parallelized.
but i heard the high IO task's limit is on IO speed.
I am not so good in hardware design.
So, can multiprocess speed up compress?
Yes, it can. Compression is usually compute bound, not I/O bound. You can use pigz to get close to a factor of n speedup for n cores.