How to choose the “number of threads” to shorten the simulation time of INTERCONNECT projects



Lumerical Solutions software utilize multi-threading which allows some calculations to be distributed to all of the cores in your computer. Multi-threading is enabled by default, since it generally increases the speed of calculations. However, this does not necessarily mean a larger “number of threads” will give a shorter simulation time. The actual simulation time depends on the combination of multiple factors such as the simulation type (frequency-domain simulation vs. time-domain simulation), the size of the simulation file, the memory type (save to disk vs. save to memory) and so on. The default “number of threads” used in INTERCONNECT is 4, which gives the best scale of simulation time on simulation size for most of the cases. However, for different cases, user has to judge differently and choose the “number of threads” accordingly.