QW-GPUSim is currently optimised for application on modern PC graphic cards of a massively parallel architecture. For the above reasons, only the parts of the code judged crucial to the speed of processing have been ported to OpenCL. Thus the gain in speed with respect to a regular CPU version depends on the actually simulated scenario and the required postprocessings.
As shown by the charts, for the QW-V2D Cassegrain antenna model, one obtains the FDTD speed of about 1600 [Mcells/s] and a speed-up of 22 when comparing GPU GTX 580 against CPU OMP I7 950. For the 3D beefburger model, one obtains the speed of about 1200 [Mcells/s] and a speed-up about 11. For the 3D antenna models with absorbing boundary conditions and NTF processing, the FDTD speed is about 750 [Mcells/s] and speed-up is equal to 10.
QW-GPUSim is a version of QW-Simulator designated for massive parallel computing hardware. It incorporates parts of QW-Simulator code re-written by QWED in OpenCL. This strategy stabilises QW-GPUSim on the very dynamic GPU/CPU hardware market.