ANSYS, Inc. Release Notes
ANSYS, Inc. Release Notes
Release 13.0 includes the following new enhancements that improve solution procedures and features.
The following enhancements are available for Distributed ANSYS:
Enhanced scalability is available in these areas:
Parallel equation ordering scheme is now default for the distributed sparse solver.
Improved scalability of assembly code for unsymmetric matrices.
Improved scalability during creation of the PCG preconditioner.
The following features are now supported and use distributed memory parallelism within Distributed ANSYS:
SURF151 and SURF153 surface elements that use the element behavior of the underlying solid element (KEYOPT(3) = 10).
Modal analyses using the unsymmetric or damped solution (MODOPT,UNSYM or MODOPT,DAMP, respectively).
Multiple load vectors and enforced motion in a modal analysis (MODCONT).
Cyclic symmetry full harmonic analyses.
Tracking nodal and element solution data (NLHIST,NSOL and NLHIST,ESOL).
Many additional features now work in Distributed ANSYS, but do not use distributed memory parallelism. For more information, see Supported Analysis Types and Supported Features in the Distributed ANSYS Guide.
It is becoming increasingly common to use the graphics processing unit (GPU) on certain high-end graphics cards to perform general-purpose computations. This capability, now available in Mechanical APDL, can accelerate portions of a simulation. Solution performance is especially improved, as the acceleration benefit applies primarily to the equation solvers, which use highly parallel, heavy number-crunching algorithms ideal for off-loading from the CPU onto the GPU.
GPU acceleration is supported on the Windows 64-bit and Linux x64 platforms. Currently, only a single GPU accelerator device can be used by the Mechanical APDL program during a solution. Only the NVIDIA Tesla GPU cards are supported, with the Tesla 20-series cards being recommended. The ACCOPTION command controls GPU acceleration options. GPU acceleration is not supported in Distributed ANSYS.
For more information, see "GPU Accelerator Capability" in the Advanced Analysis Techniques Guide.