Sun Microsystems is bringing its vision of a single binary compatible architecture to the South African market. The e-commerce player has introduced its latest clustering solution designed for resource-intensive, high performance and technical computing environments.
"It is clear that the future of supercomputing is network-centric. With our new Sun HPC ClusterTools 4 software release, we are pushing well into the Terascale realm," says Tertius Bezuidenhout, national systems engineer manager at Sun Microsystems SA.
The ClusterTools provide a complete environment to develop and manage a workload of highly resource-intensive applications on individual Sun systems, as well as a cluster of these systems.
The Sun HPC ClusterTools software also brings Sun's network computing vision to the high-end technical market. "We offer a market-leading integrated parallel application development environment that enables end-to-end scalability from individual single-processor Sun Blade workstations to the Teraflop capable clusters of our Sun Fire midframe servers," says Bezuidenhout.
The enhancements and features of the latest ClusterTools all form part of Sun's ONE (Open Net Environment) vision. "Our ONE vision is based on the integratable stack concept and strives to ensure continuous real-time. Through our latest offering we ensure elimination of downtime and little service interruption."
The source code for Sun HPC ClusterTools software is freely available for developers through the Sun Community Code License (SC) model. This model is a result of Sun's continued commitment towards the open development of key technologies.
Share
Sun HPC ClusterTools 4 software includes a new high-performance multi-protocol implementation of the industry standard Message Passing Interface (MPI), a full implementation of the MPI I/O protocol and tools for debugging, performance analysis and tuning of technical computing applications.
The software suite also features the Prism graphical programming environment, allowing developers the ability to effectively debug, performance monitor and tune large-scale parallel applications with the same efficiency as singly process applications.
Editorial contacts