by Tim Staub, associate editor LIVEwire
Dallas, Texas — Following is an interview with Steve Campbell, Director, ESP Marketing Systems Product Group, Sun Microsystems:
HPCwire: What is the latest, most exciting development now at Sun?
CAMPBELL: Scalable high-performance provides the customer the computing power they need as they need it. We have a scalable, binary server line offering four to sixty-four processors; clustering software that makes 1,024 nodes work like a single application server and network storage for data-intensive tasks. Customers are able to add memory, processing power and storage when and where they need them and rely on exceptional customer support.
HPCwire: What is the Technical Compute Farm?
CAMPBELL: The Technical Compute Farm (TCF) – a virtual supercomputing farm, acts as a single entity on the network, integrating servers, fibre-channel storage arrays, and networking with a host of supporting software components for better use of computing resources. This dramatically increases resource utilization for maximum productivity and provides high compute density per rack, at a resonable cost. TCF can be expanded from a few CPU’s to a few thousand and provides binary compatibility with next generation systems.
HPCwire: Regarding the “Net Effect”? How is Sun utilizing the Web as a resource?
CAMPBELL: HPC Portal Computing is making super-computing resources accessible through HPC.com, a highly available secure Web infrastructure. This gives businesses and research facilities platform independent thin client systems to access Web-centric HPC resources, resource pools consisting of servers, networks, storage, operating systems and advanced technical applications and libraries.
HPCwire: What processor is used to handle the “Net Effect”?
CAMPBELL: The UltraSPARC III, Sun’s second generation 64-bit architecture, was designed to meet the demands of the “Net Effect”. The UltraSPARC III is one of the most complex processors available, providing scalability, data bandwidth, advanced reliability and high availability. Guaranteed 100 percent binary compatability with the entire SPARC processor family protects the clients investment.
HPCwire: Tell us about HPC ClusterTools 3.1
CAMPBELL: HPC ClusterTools 3.1 is package of development tools that connect multiple shared memory machines for supercomputing class performance. HPC ClusterTools 3.1 is compatible with Solaris and provides enhanced debugging capabilities for MPI (Message Passing Interface). Programmers can use Prism software to debug complex programs. This makes it easier to develop highly scalable parallel applications, and support existing applications on Sun systems. ClusterTools includes: Sun Cluster RunTime Environment (CRE), Prism, Sun MPI, Sun MPI I/O, Sun Parallel File System (PFS), Sun Scientific Subroutine Library (S3L), Cluster Console Manager (CCM) and Switch Management Agent (SMA). As always, users can count on readily available, exceptional Sun support.