Run python parallel processes

Question

I have more than 20 independent scripts that I need to execute I don't know exactly how many processes I can execute in parallel as good practice:

import os
import glob
from multiprocessing import Pool

processos = (
    'p1.py',
    'p2.py',
    'p3.py',
    'p4.py',
    'p5.py',
    'p6.py',
    'p7.py',
    'p8.py',
    'p9.py',
    'p10.py',
    'p11.py',
    'p12.py',
    'p13.py',
    'p14.py',
    'p15.py',
    'p16.py',
    'p17.py',
    'p18.py',
    'p19.py',
    'p20.py',
    'p21.py',
    'p22.py',
    'p23.py',
    'p24.py'
)
def roda_processo(processo):                                                            
    os.system('python {}'.format(processo))


pool = Pool(processes=24)                                                        
pool.map(roda_processo, processos)

Would you like to know the most effective way to run these scripts.

Detail: These processes will run with schedules.

Answer 1

Why not use shell?

seq 1 24 | xargs -L1 -I{} -P`nproc` python p{}.py

 -P maxprocs
         Parallel mode: run at most maxprocs invocations of utility at
         once.

 nproc
          print the number of processing units available

Answer 2

Using GNU Parallel you can do:

parallel python ::: p*.py

It will spawn one process per CPU thread.

If your processes only use 30% CPU power (eg maybe it is waiting for the network the rest of the time) it makes sense to spawn 3 times as many:

parallel -j300% python ::: p*.py