ANSYS Mechanical HPC brings together a wide variety of high-performance computing (HPC) solutions together, including ANSYS® VT Accelerator™ and parallel computing.
VT Accelerator: Variational Technology Applied to Solver Speedup
Variational Technology has been applied to two distinct types of mathematical problems: nonlinear solutions for structural and thermal analyses and harmonic analysis. These capabilities are referred to as VT Accelerator. VT Accelerator provides a 2X to 5X speedup for the initial solutions depending on the hardware, model and type of analysis. VT Accelerator makes re-solves 3X to 10X faster for parameter changes, allowing for effective simulation-driven parametric studies of nonlinear and transient analyses in a cost-effective manner. You can make the following types of changes to the model before a VT Accelerator re-solve:
- Modify, add or remove loads (constraints may not be changed, although their value may be modified)
- Change materials and material properties
- Change section data and real constants
- Change geometry, although the mesh connectivity must remain the same (that is, the mesh must be morphed)
VT Accelerator for Nonlinear Solution Speedup
VT Accelerator for nonlinear solutions speeds up the solution of applicable nonlinear analysis types by reducing the total number of iterations. VT Accelerator supplies an advanced predictor–corrector algorithm based on Variational Technology to reduce the overall number of iterations for nonlinear static and transient analyses. It is applicable to analyses that include large deflection, hyperelasticity, viscoelasticity and creep nonlinearities. Rate-independent plasticity and nonlinear contact analyses may not show any initial improvement in convergence rates; however, you may choose this option with these nonlinearities if you wish to resolve the analysis with changes to the input parameters. In general, VT Accelerator can be used for:
- Nonlinear structural static or transient analyses not involving contact or plasticity
- Nonlinear thermal static or transient analyses
VT Accelerator for Harmonic Analysis
The harmonic sweep feature of VT Accelerator provides a high-performance solution for forced-frequency simulations in high-frequency electromagnetic problems and structural analysis. For a structural harmonic analysis, the material may have frequency-dependent elasticity or damping.
Distributed ANSYS: Parallel Power from Multi-Core to Clusters
Time is money! At ANSYS, we understand how much time means to you and that multi-processing is one means to reduce analysis time. Multi-processing computer environments (consisting of multi-processor servers or networked workstations or clusters) may be employed to generate simulation results much more quickly. The parallel portion of ANSYS Mechanical HPC facilitates this highly effective means of operation.
With Distributed ANSYS, part of the ANSYS Mechanical HPC module, the entire solution phase runs in parallel including the stiffness matrix generation, linear equation solving and results calculations. Because each of the three main parts of the overall solution are running in parallel, the wall clock time is significantly reduced. On distributed hardware, the memory required also is distributed over multiple systems. This memory-distribution method allows you to solve very large problems on a cluster of machines with limited memory. With multiple processors, you can see significant speedup in the time it takes to run your analysis for both linear and nonlinear analyses.
Revolutionary is the new Distributed PCG Lanczos solver, which allows for the quick extraction of eigenvalues and eigenmodes for very large models greater than 100 million degrees of freedom. In addition to the Distributed PCG Lanczos, Distributed ANSYS supports Distributed PCG and Distributed Sparse solvers. As always, Distributed ANSYS works equally well on both distributed memory and shared memory hardware.
Multi-Processing Computer Environments
- Windows® CCS
- Windows x64 clusters
- 64-bit Linux® clusters
- 32-bit Linux clusters
- UNIX® servers
- Networked workstations
Memory Architectures Supported
- Shared memory
- Distributed memory
- Mixed memory
Analysis Types Supported by Distributed ANSYS
- Static linear or nonlinear analyses
- Full transient structural analyses
- Static thermal analyses
- Full transient thermal analysis
- Magnetostatic, steady-state electric conduction and electrostatic analyses
- Time-harmonic and time-transient magnetic and electric analyses