Shared Memory & Distributed Memory Systems
To provide faster turn-around, problems can be run in parallel on multiple cores, processors or clustered computer systems. Domain decomposition is used to produce the most efficient parallel solution. Simulations can run in parallel on two processors with the standard installation, without the need for additional message-passing software or HPC licenses. In most cases, using two processors will result in a speed improvement by a factor very close to two. (Users of ANSYS Autodyn single task are required to obtain HPC licenses).
 |
|
Close to linear speedup obtained with parallel processing
|