Python Code Parallelization For Alma, Challenges And Alternatives
Justo Antonio González Villalba, Julian Taylor, S. Castro, Jeffrey S. Kern, J. Knudstrup, Stefano Zampieri, Alisdair Manning · Zenodo (CERN European Organization for Nuclear Research) · 2015
In the last few years development of Python code for science and data reduction purposes has gained significant popularity. ESO in itself uses a Python-based archiving system for VLT and ALMA data. Also the data reduction suite for ALMA data is python-based. Rapid development is fostered by a big community and a wide range of already available packages. However Python enforces locking mechanisms, to ensure thread safety, that effectively reduce the capacity of Python to use only one core. In this context a number of alternatives have been developed by the community to emulate actual multi-threading and make parallel processing easier to use from Python, preserving interactivity.