Performance Computing

High-performance computing (HPC) is an important consideration for academia. Complex research projects often involve simulating large assemblies, even entire systems which in turn demand considerable computational resources when rapid results are required. The widespread availability of multicore processors, even in laptops, allows for impressive performance gains for medium to large numerical problem sizes. The ANSYS Academic products meet these HPC requirements by providing two levels of HPC:

  Built-in HPC:                 (-)   

For Teaching, Research & Associate level licenses HPC capability is built into each task of the Academic base licenses. For most physics solvers this is both shared & distributed memory HPC capability for up to four cores, plus GPU access is enabled for ANSYS Mechanical & ANSYS Fluent solvers.

All of the ANSYS Academic Products that bundle solvers that support HPC have four core HPC capabilities built-in. This means that up to 4 CPU cores can be used to accelerate simulation solution times. For the structural, thermal & fluids academic products:

  • Built-in HPC capability is allocated/restricted to each task. For example, for a 25-task license of ANSYS Academic Research, each user/task can access up to four cores, with a maximum of 4 cores available per solution.
  • Shared Memory (SMP), Distributed Memory Parallel (DMP) and Domain Decomposition methods are included.
  • GPU access is enabled for ANSYS Mechanical & ANSYS Fluent solvers, allowing multiple GPU cards to be used. We do not restrict the number of GPU cards for the academic products.
  • Frequency Sweep VT & VT Accelerator is built-in and available for the ANSYS Mechanical APDL solver. This is a legacy HPC frequency sweep solution that has been largely replaced by DesignXplorer & Design Point Updates.
  • The following structural, thermal & fluids solvers support HPC:
  • ANSYS CFX (SMP, DMP)
  • ANSYS Fluent (SMP, DMP)
  • ANSYS Mechanical (SMP, DMP, Domain Decomposition)
  • ANSYS Autodyn (SMP, DMP)
  • ANSYS Polyflow (SMP, DMP)
  • ANSYS Icepak (SMP, DMP)

For the electromagnetics academic products:

  • Built-in HPC capability is not allocated/restricted to each task. For example, for a 25-task license of ANSYS Academic Research HF, one hundred HPC tasks are built-in, so a single user can use all 100 HPC tasks, or alternatively 25 users can each use 4 cores per solution.
  • A legacy shared memory (SMP) multiprocessor capability is enabled, this can make use of any number of SMP cores, but does not scale well beyond 4 cores. It has been largely superseded by other HPC methods.
  • Distributed Memory Parallel (DMP) and Domain Decomposition methods are included.
  • The following electromagnetics solvers support HPC:
  • ANSYS HFSS (SMP, Domain Decomposition, Distributed Frequency Sweep)
  • ANSYS Maxwell (SMP, Distributed Frequency Sweep)
  • ANSYS Designer (SMP, Distributed Frequency Sweep)
  • ANSYS Q3D (SMP, DMP)
  • ANSYS SIwave (SMP)
  • ANSYS TPA (MP, Distributed PKG analysis)
  Add-on HPC:                 (-)   

For Research & Associate level licenses, the built-in HPC capability can be extended by purchasing additional ANSYS Academic HPC licenses. For most physics solvers both shared & distributed memory HPC capability can be extended, plus GPU access is extended for ANSYS Mechanical & ANSYS Fluent solvers.

Add-on Academic HPC Product Extends the built-in HPC of the following base product(s)
ANSYS Academic Associate HPC ANSYS Academic Associate
ANSYS Academic Associate CFD
ANSYS Academic Research HPC ANSYS Academic Research
ANSYS Academic Research CFD
ANSYS Academic Research Mechanical
ANSYS Academic Research Autodyn
ANSYS Academic Research Polyflow
ANSYS Academic Research Electronics Thermal
ANSYS Academic Associate Electronics HPC ANSYS Academic Associate EM
ANSYS Academic Associate HF
ANSYS Academic Research Electronics HPC ANSYS Academic Research EM
ANSYS Academic Research HF