ANSYS Electronics HPC enables parallel processing for solving the toughest and most challenging models — models with great geometric detail, large systems and complex physics. ANSYS goes well beyond simple hardware acceleration to deliver groundbreaking numerical solvers and HPC methodologies optimized for multicore machines, with scalability to take advantage of full compute cluster power. The amount of HPC required is based simply on the total number of cores used in the analysis, irrespective of which HPC technology is employed.
Multithreading: ANSYS Electronics HPC takes advantage of multiple cores on a single computer to reduce solution time. Multithreading technology speeds up the initial mesh generation, matrix solves and field recovery.
Spectral Decomposition Method: The spectral decomposition method (SDM) accelerates frequency sweeps by distributing multiple frequency points in parallel over compute cores and nodes. You can use this method in tandem with multithreading to speed up extraction of individual frequency points, while SDM parallelizes multi-frequency point extraction.
Domain Decomposition Method: The domain decomposition method (DDM) accelerates the solution for larger and more complex geometries by distributing a simulation across multiple cores and networked nodes. This method is primarily designed to tackle larger problem size using distributed memory. It can also be combined with multithreading and SDM to provide improvements in simulation scalability and throughput.
Periodic Domain Decomposition: Periodic domain decomposition applies DDM to finite periodic structures such as antenna arrays or frequency selective surfaces. This method virtually duplicates the geometry and mesh of the periodic structure’s unit cell and then applies the DDM algorithm to the resulting finite sized array to solve for the unique fields for all elements. Simulation capacity and speed are substantially increased. This method can be combined with multithreading and SDM to further accelerate the solution.
Hybrid Domain Decomposition Method: Hybrid DDM uses the domain decomposition method on models consisting of finite element (FE) and integral equation (IE) domains. The HFSS IE solver add-on lets you create HFSS models that can solve extremely large EM problems. This methodology combines the virtues of FEM’s ability to handle complex geometries plus MoM’s efficient solutions for antenna and radar cross section analysis. Hybrid DDM can be combined with multithreading and SDM to provide further solution acceleration.
Distributed Direct Matrix Solver: The distributed direct matrix solver is a distributed memory parallel technique for HFSS and the HFSS-IE solvers. The matrix solution is distributed across multiple cores or MPI-integrated computers. It results in solutions with improved scalability through increased MPI memory access, and enhanced speed through increased MPI networked core access for highly accurate direct matrix solver solutions. These distributed memory matrix solvers can be combined with multithreading and SDM to further increase simulation throughput.
Distributed Memory Matrix Solver: The distributed memory matrix solver (DMM) is a distributed memory parallel technique for HFSS, including the finite element method (FEM) and integral equations (IE). The matrix solution is distributed across multiple cores of MPI-integrated compute nodes. It results in a reduced memory footprint per node and improves scalability and speed through increased MPI memory access and networking. The DMM solver is integrated in the Auto-HPC technology and can be orthogonally combined with the spectral decomposition method (SDM) to further increase simulation throughput.
HPC in the Cloud: The ANSYS Cloud service makes high-performance computing (HPC) extremely easy to access and use. It was developed in collaboration with Microsoft Azure, a leading cloud platform for HPC. It has been integrated into the Electronics Desktop, so you can access unlimited, on-demand compute power from the design environment.