HLA-Based Parallel Simulation: A Case Study

As a distributed simulation standard, the high level architecture (HLA) is used to facilitate the interoperability and reuse of multiple federates, rarely applied in the high-performance parallel simulation. This paper, however, shows a parallel simulation running on a 10 teraflops cluster system, which is based on the master-slave programming model, developed in the use of runtime infrastructure (RTI). After the simulation system is introduced, the framework of its core program is also presented as well as the key techniques of object distribution, event management and time management in the program. The HLA-based simulation supports the single program multiple data (SPMD) execution mode and runs well presently in the high-performance parallel computer.