GPU-based parallelization of system modeling

Simulation of fiber-optic transmission systems using general purpose-computing on graphics processing units is investigated. It is shown that a speedup factor of more than 100 can be achieved compared to CPUs without loss in accuracy.