Provide modular, open-source and flexible analysis toolkit components for associated workflows to enable rapid, reproducible and scalable analysis of the large-volume and complex data products created by current, and future, radio and millimetre astronomical research infrastructures (RIs). These toolkit components will enable user exploitation of innovative techniques (such as AI/ML and Bayesian inference), translate knowledge and developments between different RIs; and provide RI science users with the elements of workflow implementations that will allow them to exploit, and contribute to, cutting-edge industry and academic-driven high-performance compute advances such as GPU, tensorcore, distributed compute and cloud environments.
Results | Link |
---|---|
D5.1 Library of DASK-accelerated interface to analyse radio astronomy data formats | https://zenodo.org/records/10940716 |
D5.2 Implementation of a fringe fit algorithm in the Dask framework | |
D5.3 Prototype processing workflow functionality using software modules under DASK framework | |
D5.4 Port of Bayesian package to make use of DASK-like HPC methods | |
D5.5 Demonstrated usage if end-to-end simulation tool for single R1 | |
D5.6 Full-Stokes dynamical imaging algorithm for time-variable VLBI sources | |
D5.7 Software/firmware repositories containing developed code and documentation | |
D5.8 DASK accelerated data reading & GPU processing for radio astronomy tools |