Different implementations of parallel processing for FDTD simulator

Parallel processing in FDTD as well as other electromagnetic simulation methods is growing in importance. It allows not only to calculate faster but also to cope with large structures. Parallel processing can be implemented in different ways. The implementation depends mainly on computer architecture and programming libraries under different operating systems. Usually, different approach has to be taken also for a newly created program in contrary to rebuilding existing serial program into a parallel version. A European project for development of a parallel version of existing, professional FDTD simulator is presented. The simulator was principally designated for serial execution. Consecutive steps of code transition from serial into versatile parallel version are described.