This module contains decorators for parallel vectorization of functions.

The decorators vectorize the function over the first argument and return a list of the returned results of the functions. The main decorator is vectorize_parallel that supports parallelization via the multiprocessing module or via MPI.

• Example for multiprocessing:

   from parallel_decorators import vectorize_parallel
   @vectorize_parallel(method='processes', num_procs=2)
   def power(x, y):
       return x**y
   result = power(range(5), 3)
   # result is now: [0, 1, 8, 27, 64]
  

• Example for MPI:

   from parallel_decorators import vectorize_parallel, is_master
   @vectorize_parallel(method='MPI')
   def power(x, y):
       return x**y
   # computation in parallel here
   result = power(range(5), 3)
   # result is now: [0, 1, 8, 27, 64]
   if is_master();
      # use results on master core
      print(result)
  

  Then start script with

   $ mpiexec -np <num> python script.py
  

The module can be installed using setup.py by calling

python setup.py install

or if you do not have access to the system location of python

python setup.py install --user

Afterwards, the module can be imported and used.

If you want to use the MPI backend, you have to install the mpi4py package (e.g. with pip install mpi4py --user).

If you want to use the progressbar feature, you have to install the progressbar2 package (e.g. with pip install progressbar2 --user).