Interactive in-situ steering by file-base control using particle-based in-situ visualization framework

In a large-scale simulation on supercomputer, it is difficult to repeatedly perform recalculation due to adjustment of computational parameters. We have developed an interactive steering method for the simulation parameter scan by file-based control on an in-situ visualization framework using particle-based volume rendering. This framework was applied to plume dispersion simulation in urban area on GPU cluster, and the steering for pollution source position was implemented. Through human-in-the-loop interactive steering, we succeeded in searching for pollution source locations that meet the pollutant concentration conditions at observation points in the urban area. (author)