The high complexity of the systems and the strict requirements with respect to response times and stability in a multi user environment there has increased the demand of high-performance computing (HPC). Parallel computing has become the dominant paradigm in computer architecture.

High-performance computing uses Computer Cluster , which may be distributed (Grid Computing). Also other computer systems with very large and efficiently organized storage capabilities can be used for high-performance computing.

In order to profit from this computer architecture the applications must be designed using parallel programming models. This may be achieved on a macro level (parallel processing on several computers) or on a micro level (Multi Core Programming).

Grid Computing

Grid computing is the application of distributed and autonomous resources. Grid or distributed computing is an intensive field of research, the vision is the use of IT resources of all kinds. Nets, computers and programming languages must all be compatible and thus lead to a new dimension of virtual organization. The term grid has its origin from the electricity grid. So the IT power shall be transmitted through the grid as easily and naturally as electricity.

In the financial sphere parallelism and speed mean a big challenge for the front and back office systems. Complex computations in mathematical finance from valuation and optimisation to the quantification of market and credit risks are unavoidable. In many cases simulation methods are used which in turn is a challenge for real time processing.

In addition to the intellixx solution (ifs.grid) we also support commercial solutions of our technology partners. These solutions are ready for use and supported by most financial systems of our implementation partners. We have  active partnerships with respect to

• DataSynapse
• Platform Company
• Projekt FinGrid

Multi Core Programming

Standard software, usually written for a single processor, does not necessarily run faster on a multi core processor architecture. The relevant parts of the software must be coded for parallel processing and must be multi threaded in order to take advantage this architecture.