HPC for Fluids
Parallel processing with ANSYS High Performance Computing (HPC) enables you to solve large models with the most complex CFD operations faster and at lower cost. It is optimized to run our solvers, such as ANSYS CFX and Fluent, on the latest HPC architectures, like Haswell processors and NVIDIA GPUs, as well as Intel True Scale and Cray MPI (interconnects). ANSYS HPC has superior and proven scalability — it operates at 80 percent efficiency with as low as 5,000 cells per CPU core, and can enable extremely large simulations, 10,000+ cores for models with 100 million cells or more.
- Reduces time to solution
- Increases high-fidelity insight
- Increases number of design variants
- ANSYS HPC has superior and proven parallel scalability above 80 percent efficiency with as low as 5,000 cells per CPU core. This means that:
- You will be less constrained by hardware limitations because bigger models can be sped up at using lesssmaller HPC hardware, – at lower cost.
- You can benefit from ground-breaking scale-up performance at very high core counts for large models – e.g., 15,000+ cores for more than 100 million+ cells.
- ANSYS solvers required for complex physics (chemistry, multiphase) are highly optimized to run fast and deliver outstanding parallel scaling on today’s multicore processors.
- Besides parallel scalability, solver convergence and robustness are of paramount importance. ANSYS CFX and Fluent are highly valued for these characteristics.
- Our solvers are optimized to run on the latest HPC architectures, like processors NVIDIA GPUs, and Intel True Scale (interconnect) fabrics. For example, GPU support for the 3D-AMG coupled pressure-based solver in Fluent demonstrates the ANSYS’ commitment to allowing customers to leverage new and evolving technology, such as GPU, for faster simulation.
- Fluent meshing leverages its unmatched HPC capabilities and scalability to deliver parallel meshing, which allows enables dramatic reduction in meshing time. For example, using eight cores, a 42 million cell mesh can be meshed more than seven times faster than if a single CPU core was used. Note that this capability does not require HPC licenses!
- ANSYS provides flexible and scalable licensing.