.. _parallel:
========
Parallel
========
Installation
------------
Running this code in parallel requires the installation of `MPI for Python
`_ which requires that your system has mpi
installed.
Running in parallel
-------------------
Depending on what modules, methods, and/or functions from BET your script uses
small or no changes will need to be made to run your script in parallel. To run
your script in parallel you will have to use::
$ mpirun -np NUM_PROCS python YOUR_SCRIPT.py
instead of ::
$ python YOUR_SCRIPT.py
Example scripts are availiable in ``examples/parallel_and_serial_sampling``
which demonstrate different modes of parallel and serial sampling. Also the
:ref:`fenicsMultipleSerialExample` is an example that can be run with serial
BET and uses `Launcher `_ to launch multiple
serial runs of the model in parallel.
Parallel Enabled Modules
------------------------
The modules that have parallel capabilities are as follows::
bet/
util
sample
Comm
surrogates
calculateP/
calculateP
simpleFunP
calculateError
sampling/
basicSampling
postProcess/
plotP
postTools
sensitivity/
chooseQoIs
util
~~~~
The module :mod:`~bet.util` provides the method
:meth:`~bet.util.get_global_values` to globalize local arrays into an array of
global values on all processors.
sample
~~~~~~
The :class:`~bet.sample.sample_set_base` has methods which benifit from
parallel execution and has methods for localizing
(:meth:`~bet.sample.sample_set_base.local_to_local`) and localizing
(:meth:`~bet.sample.sample_set_base.local_to_global`) member attributes.
Localized attributes are denoted as ``*_local``. The
:class:`~bet.sample.discretization` aslo has methods which enable and benifit
from parallel execution.
calculateP
~~~~~~~~~~
All methods in the module :mod:`~bet.calculateP.calculateP` benefit from
parallel execution.
In the module :mod:`~bet.calculateP.simpleFunP` any method with a
``num_d_emulate`` option benefits greatly from parallel execution.
sampling
~~~~~~~~
If you are using a model with parallel capabilities we recommend that you write
your own python interface to handle running multiple parallel copies of your
model simultaneously. If your model is serial then you might benefit from
parallel execution of scripts that use
:class:`bet.sampling.basicSampling.sampler`. The method
:meth:`~bet.sampling.basicSampling.sampler.compute_qoi_and_create_discretization`
and :meth:`~bet.sampling.basicSampling.sampler.create_random_discretization`
both will partition the samples over several processors and have a globalize
option to return a globalized set of results.
postProcess
~~~~~~~~~~~
In :mod:`~bet.postProcess.plotP` the methods
:meth:`~bet.postProcess.plotP.calculate_1D_marginal_probs` and
:meth:`~bet.postProcess.plotP.calculate_2D_marginal_probs` benefit from
parallel execution. The methods :meth:`~bet.postProcess.plotP.plot_1D_marginal_probs` and
:meth:`~bet.postProcess.plotP.plot_2D_marginal_probs` will only execute on the
rank 0 processor.
In :mod:`~bet.postProcess.postTools` the methods
:meth:`~bet.postProcess.postTools.save_parallel_probs_csv`,
:meth:`~bet.postProcess.postTools.collect_parallel_probs_csv`,
:meth:`~bet.postProcess.postTools.save_parallel_probs_mat`, and
:meth:`~bet.postProcess.postTools.collect_parallel_probs_mat` provide tools to
save and collect probabilities on separate processors as appropriately named files.
sensitivity
~~~~~~~~~~~
All methods in the module :mod:`~bet.sensitivity.chooseQoIs` benefit from parallel execution.