Nasa takes advantage of parallel computing
Nasa is to use Interactive Supercomputing’s Star-P software to take advantage of the huge speed and memory capability of high-performance computers in the next generation of space telescopes and stellar imaging applications.
ISC has won a grant to link Star-P to Nasa’s Optical Systems Characterisation and Analysis Research (Oscar) software. It will allow scientists to reuse existing serial and parallel Oscar code, without having to learn Oscar’s complex programming requirements.
Oscar is modelling software used to design and analyse large space-based imaging systems, and was instrumental in solving the well-known optical flaws in the Hubble Space Telescope. Due to its demanding processing and memory requirements, Nasa runs Oscar on Beowulf parallel computing clusters.
Using Star-P, scientists can reuse the existing code by linking it to easy-to-use tools such as Python or Matlab. Star-P will then automatically transform the programs tp run on the Beowulf clusters.
‘While OSCAR is extremely powerful, it is not easy to use since there is no GUI and it requires calling library functions from C or MPI code,’ said Alan Edelman, ISC’s chief science officer. ‘Interfacing OSCAR with Star-P will enable scientists to breathe new life into old code using easy-to-use, high-productivity desktop environments, without having to re-program the applications in C or MPI to run on parallel systems.’