June 29th, 2010
PyMW 0.4 has been released after work to update the interfaces and add support for Python 3.
Major Changes in 0.4
- Added support for Python 3.0 (tested with Python 3.1)
- Added Ganga interface (contributed by Wayne San)
- Added helper methods to simplify application setup
- Improved error detection and handling
- Changed MPI interface to use mpi4py (included with SciPy) rather than pympi
PyMW 0.4 (June 29, 2010)
September 20th, 2009
PyMW 0.3 has been released after work over the summer as part of Google Summer of Code.
Changes in 0.3
- Added MapReduce class for use with PyMW.
- Added interface for performing simulations.
- Added worker-task matching functionality.
- Added support for file attachments to tasks.
- Added traceback for worker exceptions.
- Simplified interface design and function.
- BOINC interface changes
- Added native worker for launching scripts in Windows.
- Added support for custom, cross-platform workers.
- Removed old pymw_setup script (automatic now).
- Added attributes to customize BOINC work units.
- Updated to use the new structure for BOINC Python library.
PyMW 0.3 (September 20, 2009)
June 24th, 2009
We are in the process of setting up WordPress for our project page. New PyMW content and documentation are on the way!
In the mean time, here is a little information about PyMW:
PyMW is a Python module for master worker style parallel computing in a variety of environments. Condor, and BOINC.
Pydoc generated documentation is available here.
A technical report with more detail is available here.
PyMW was developed by Eric Heien. The BOINC interface code was developed by Adam Kornafeld and Jeremy Cowles. To report bugs or feature requests, please use the Sourceforge interface. PyMW is released under the MIT License. See the README file in the PyMW distribution for more information