Currently saber is written such that the user needs to set up a lot of the multiprocessing in the script which executes all the functions. Can the multiprocessing get moved upstream inside the corresponding saber functions? that could lead to repeating the overhead to create a multiprocessing pool but may still be clearer/easier to use while still being nearly as efficient.
Currently
saberis written such that the user needs to set up a lot of the multiprocessing in the script which executes all the functions. Can the multiprocessing get moved upstream inside the corresponding saber functions? that could lead to repeating the overhead to create a multiprocessing pool but may still be clearer/easier to use while still being nearly as efficient.