A common interface for parallel processing pools.
A common interface to parallel processing pools.
Enables easy switching between serial execution, multiprocessing, and MPI by providing a common interface.
If you can write your code following the MapReduce programming model, you can trivially parallelize it using The Schwimmbad. For example, if you can structure your code so that a ‘worker’ function operates on elements from a list of tasks, like this:
def worker(task): # do something with the task! return task**2 def main(): tasks = range(100) results = map(worker, tasks) if __name__ == "__main__": main()
with a few added lines of code, you can make it runnable with multiprocessing and MPI at the switch of a command-line flag:
import schwimmbad def worker(task): # do something with the task! return task**2 def main(pool): tasks = range(100) results = pool.map(worker, tasks) pool.close() if __name__ == "__main__": from argparse import ArgumentParser parser = ArgumentParser(description="") group = parser.add_mutually_exclusive_group() group.add_argument("--ncores", dest="n_cores", default=1, type=int, help="Number of processes (uses multiprocessing).") group.add_argument("--mpi", dest="mpi", default=False, action="store_true", help="Run with MPI.") args = parser.parse_args() pool = schwimmbad.choose_pool(mpi=args.mpi, processes=args.n_cores) main(pool)
The easiest way to install is via pip:
pip install schwimmbad
The project is also installable by cloning or downloading the source and using the standard:
python setup.py install
Copyright 2016 Adrian Price-Whelan.
schwimmbad is free software made available under the MIT License. For details see the LICENSE file.