9 January 2008
Interactive Supercomputing (ISC) has introduced a new version of its Star-P software for parallel programming, with increased support from different programming languages, operating systems and workload management platforms.
Star-P 2.6 features a preview version of an R language client, expanding its support for Very High Level Languages (VHLL) beyond Matlab and Python. The new version also extends Star-P’s support for Python, with more than 70 new parallel computing functions added to its Python mathematical libraries and improved file I/O. In addition, the software is now compatible with SciPy, an open source library of Python programming extensions that support large, multi-dimensional arrays and matrices.
ISC has also increased its software compatibility with different cluster workload managers. In addition to PBSpro, Star-P can now integrate with Sun Microsystems’s Sun Grid Engine.
The software now has the ability to load mathematical libraries that are optimised for a specific computing platform. Previous versions were not hardware specific, requiring additional fine-tuning of the libraries to optimise performance. The new version automatically recognises and loads mathematical libraries pre-optimised for a specific hardware platform, resulting in performance gains of 20 to 100 per cent.