Time is Money! At ANSYS, Inc., we understand how much time means to you and understand that multiprocessing is one means to reduce analysis time. Multiprocessing computer environments (consisting of multiprocessor servers or networked workstations or clusters) may be employed to generate simulation and analysis results much more quickly. The Parallel Performance for ANSYS add-on module facilitates this highly effective means of operation.
This add-on module features three solvers that permit engineers to exploit inexpensive cluster computer systems using distributed parallel processing and a fourth parallel solver which offers improved iterative solver performance on many difficult large and complex engineering problems.
The solvers are:
Parallel Performance for ANSYS is ready today for both your 32-bit and 64-bit systems. Leveraging your existing investment in 32-bit hardware, Parallel Performance for ANSYS can reduce the solution time due to improved scalability while increasing the maximum problem size that can be solved. By utilizing an efficient memory scheme available for tetrahedral elements, which is typical when working with complex CAD solid models that are automatically meshed, up to 8 million Degrees of Freedom can be solved on Intel IA32 Linux and Windows machines.
Allowing you to expand beyond the limits of 32-bit computing, Parallel Performance for ANSYS supports a wide variety of 64-bit systems. Combining the DPCG solver with the memory-saving option mentioned above, allows for the solution of enormous problems!
Distributed Pre-conditioned Conjugate Gradient (DPCG)
DPCG is based on a new implementation of the classic PCG solver which targets
distributed parallel processing. The DPCG solver preserves all of the robustness
of the proprietary pre-conditioner used in the PCG solver and can be run on
either shared memory or distributed memory machines with superior scalability
to the PCG solver. Compared to the DDS solver, the DPCG solver is more robust
and uses less memory with similar scalability at low (less than 16) numbers
of processors.
Distributed Jacobi Conjugate Gradient (DJCG)
DJCG is based on a new implementation of the classic JCG solver, which targets distributed parallel processing for field type problems such as thermal analysis. The DJCG supports both shared memory and distributed memory architectures. Scalability of this solver is superior to the JCG solver with little additional memory required.
Algebraic Multigrid (AMG)
The Algebraic Multigrid (AMG) Solver, which uses a sophisticated multi-level preconditioner with specific enhancements for solving ill-conditioned matrices, It is intended for use on shared-memory, multiprocessing servers. It scales well for up to eight processors, yet maintains excellent performance levels even on a single processor. In addition, the AMG solver works well with models typically difficult for other types of iterative solvers. These include ill-conditioned models resulting from large element aspect ratios or models with shell and/or beam elements attached to solid elements. The AMG solver uses a global stiffness matrix procedure for pre-conditioning, as opposed to the elemental-based procedure used by our popular PCG solver. This makes it more robust than even the PCG solver for resolving ill-conditioned problems. However, the AMG solver requires more computer memory than the PCG solver. The AMG solver is intended for linear and nonlinear, static, and full transient structural analyses.
Distributed Domain Solver (DDS)
The Distributed Domain Solver (DDS) is a scalable solver designed for use in a distributed-memory environment or a combination of both distributed- and shared-memory environments. It is intended for large static or full transient analyses with symmetric matrices that do not involve pre-stress, inertia relief, coupling or constraint equations. DDS divides a model into multiple "domains," that is, contiguous groupings of finite elements drawn from the complete model. Each individual domain is generally 1,000 to 10,000 degrees of freedom in size.
