| Feature | Available in Parallel? | Limitation | |
|---|---|---|---|
| Solvers | Structured | Yes | |
| Unstructured | Yes | ||
| SPH | Yes | ||
| NBS Tet | Yes | ||
| Rigid Bodies | Yes | ||
| Euler Ideal Gas | Yes | ||
| Euler Multi-Material | Yes | ||
| Contact | Gap | Yes | Edge on edge impact is not supported. |
| Trajectory | Yes | Analyses using trajectory contact cannot be run in parallel if the Decomposition Response method is selected. | |
| Connections | Bonded/Breakable | Yes | |
| Spotwelds | Yes | ||
| Reinforcement Beams | Yes | ||
| Joins | Structured/Unstructured | Yes |
Although the joins between unstructured parts have been parallelized it is strongly advised to avoid the use of unstructured parts in combination with joins in parallel analyses. An option is available that will merge joined unstructured nodes that reside at the same physical location in the model into one single unstructured node. The option is available under the Join menu and will increase robustness in many applications involving joins. |
| SPH-Structured | Yes | ||
| Euler Ideal Gas | Yes | ||
| Euler Multi-Material | Yes | Joins cannot be used in combination with Euler/Lagrange coupling. | |
| Euler/Lagrange Interaction | Euler Ideal Gas | Yes | |
| Euler Multi-Material | Yes | Joined Euler cannot be used in combination with Euler/Lagrange coupling. | |
| Boundary Conditions |
Stress/Velocity/Bending Flow/Transmit | Yes | |
| Analytical Blast | No | ||
| Remote Displacement | Yes | ||
| Materials | Autodyn Materials | Yes | |
| Sesame Library | No | ||
| Gauges | Fixed gauges in structured Lagrange parts | No | |
| Analysis Settings | Automatic Mass Scaling Update Frequency | No | |
Notes on Parallel Performance
Simulation time is generally expected to decrease when a simulation is run distributed across more cores, but this only occurs up until a certain number of cores. The optimal number of cores to run over is dependent on many factors, including the simulation itself (number of degrees of freedom or features used) and the hardware the simulation is being run on. The advice in Solution and Performance Issues also largely applies to Autodyn and Explicit Dynamics simulations, particularly the section on Hardware.
Note that running an Autodyn or Explicit Dynamics simulation over N cores actually requires the spawning of N+1 processes (the master process and N worker processors). Therefore, take special care not to oversubscribe any particular hardware by launching more processes on that hardware than physical cores, as this may result in poor performance.
Oversubscription is not recommended, but if it is necessary and Intel MPI is being used, there are some environment variables that may be set to improve the performance. These include setting I_MPI_PIN to OFF, and I_MPI_WAIT_MODE to 1 with I_MPI_THREAD_YIELD to 3. See the Intel MPI documentation for further details on these environment variables.