mpi-map 2.6

Tools running parallel processes

Documentation

This package uses mpi4py and dill to spawn processes and execute them.

Installation

You need to have an MPI back-end installed on your machine and add the right path on the $LD_LIBRARY_PATH, so that mpi4py can link to it. You should install mpi4py manually by

$ pip install mpi4py

When everything is set, you can install the mpi_map using:

$ pip install mpi_map

ChangeLog

v1.0.1

• map methods of classes

v1.0.3

• added export of cwd in order to allow import of local modules

v1.0.17

• Removed disconnect on stop(). This was causing an unnecessary deadlock.

v1.0.18

• Reverted change of 1.0.17. The problem with disconnect seems to be related to mpi4py.

v1.0.19

• Removed disconnect from children. Added free on master.

v1.0.20

• Allowing for the running of method belonging to an object in the distributed memory.

v2.0

• Advanced management of distributed memory

v2.1

• Fixed a leftover bug on the allocation of memory thorugh function evaluation.

v2.2

• Added barrier synchronization to prevent buffer overloading in heavily asynchronous applications

v2.3

• Added barrier sync in binary communications

v2.4

• Fixed handling of non-tuple outputs in mpi_map_alloc_dmem