Supercomputing with VPX?
Supercomputers are the top echelon of computing. They are used in the most demanding of data processing applications. Building one has always been left to the expertise of the IBMs and Crays of the the world. The advent of the microprocessor started the shift to processor complexes that could be tightly coupled through high speed network fabrics to do some of the same data processing. Specifications developed over that past few years that define computer blades with many network fabric connections per module make it possible to build some very complex data processing systems.
The members of VITA have seen the opportunity with VPX to define specific OpenVPX profiles that provide guidance in constructing special case hypercube configurations that can deliver very deterministic data processing systems. Getting ahead of the curve to define both copper and optical configurations can make it easier to build supercomputing class systems with VPX. Defining these profiles now will make it easier for the industry to come together in building the complex systems needed to execute new algorithms that are emerging.
It is going be be interesting to see how new solutions emerge out of this effort. Stay tuned for more as this working group moves forward.