par_parallel

PAR: a PARallel and distributed job crusher


License
GPL-2.0+
Install
pip install par_parallel==1.0

Documentation

If you use this software, please cite the corresponding publication:

@article{Berenger2010,
author = {Berenger, Francois and Coti, Camille and Zhang, Kam Y. J.},
title = {{PAR: A PARallel And Distributed Job Crusher}},
doi = {10.1093/bioinformatics/btq542},
year = {2010},
journal = {Bioinformatics}
}

# you need to install Pyro, cf. http://pyro.sourceforge.net
# (sudo yum install pyro _OR_ sudo aptitude install pyro)
# On the server, TCP port 7766 must be open
#                TCP port 9090 must also be open if the server is also
#                the Pyro nameserver (default case)

# if Pyro is not installed system wide, try this:
export PYTHONPATH=~/usr/Pyro-3.10/build/lib

# Some usage examples:

# run in parallel commands from test_parallel.input
cat test_parallel.input | ./parallel.py -i /dev/stdin

# same than before but with a progress bar
cat test_parallel.input | ./parallel.py -i /dev/stdin -v

# run in parallel commands from test_parallel.input and store output
# in output.log
./parallel.py -i test_parallel.input -o output.log

# same than before but with a user defined output printer
./parallel.py -i test_parallel.input -o output.log -p post_proc_example

## real world usage example
# 1) server side
./parallel.py -v -i many_commands.sh -o par_many_commands.log -s
# 2) client side, on each machine you want to join the computation
#    replace SERVER_NAME by the machine name from where you launched
#    parallel.py using -s
./parallel.py -c SERVER_NAME
# 3) be thrilled! ;)