Skip to main content

MPP Scalability

MPP (short for Massively Parallel Processing) is a type of computing available for LS-DYNA that uses many separate CPUs running in parallel each with their own memory to execute a single analysis. In order to solve large implicit CFD analyzes, it is important to provide a good CPU scalability in order to accelerate the analysis and save some computational time. A numerical model of around 2 million degree of freedom was run as a stand alone CFD model. An FSI analysis was also performed on the same model bringing the total number of degrees of freedom to around 3.5 millions. Bearing in mind that the ICFD solver is implicit, the results show good speedup capabilities : of 40 for 128 cpus in the CFD only case and a speedup of 55 for 128 cpus in the case of FSI. For the next development cycle further improvments will be implemented.