As you know, there is an option to specify the number of parallel streams when using
eos cp as shown below quoted from
eos cp --help:
Usage: cp [--async] [--atomic] [--rate=<rate>] [--streams=<n>] [--depth=<d>] [--checksum] [--no-overwrite|-k] [--preserve|-p] [--recursive|-r|-R] [-s|--silent] [-a] [-n] [-S] [-d[=][<lvl>] <src> <dst> --streams : use <#> parallel streams
However it does not affect at all the transfer performance as it increases (tried the number of streams only from 1 to 5 for just testing).
I was just wondering what exactly the number of parallel streams means here and how it runs. Also it would be very grateful to know in which case it is most effective.