High-Performance Computing Features
High-performance computing (HPC) is strategic technology that provides added value to an investment in ANSYS simulation technology.
Whatever the simulation need, ANSYS HPC provides the parallel processing capability required to accelerate time-to-solution and solve the highest fidelity problems. The portfolio of ANSYS mechanical and fluid solvers — including ANSYS Mechanical, ANSYS Autodyn, ANSYS Fluent, ANSYS CFX, ANSYS Icepak, and ANSYS Polyflow software — all use the same ANSYS HPC licenses to execute in parallel. Buy once, deploy once and use parallel processing when and where it is needed. HPC capabilities are available for ANSYS electromagnetics products as well.
ANSYS HPC is tuned for optimum performance on the latest multi-core processors from Intel® and AMD. Visit ANSYS published benchmarks and learn about 80 to 90 percent efficiency, or better, for single processor performance on a range of simulation workloads. ANSYS and ANSYS partners focus on issues like compiler optimization, process affinity setting, multi-core-aware partitioning and cache tuning in order to ensure peak performance on each new generation of processors.
For example, with the release of the new Intel E5-2600 v3 processors, ANSYS Mechanical 16.0 users can benefit from significant performance gains being realized through Intel's adoption of the latest AVX-2 compiler instructions and math kernel libraries.
ANSYS HPC delivers outstanding scaling. The largest fluids benchmarks for software from ANSYS yield nearly linear scaling up to 36,000 processor cores. This scalability reduces turnaround time from hours or days to just minutes. Mechanical simulations can achieve teraflop compute rates on hundreds of processors. Depending on problem size, ANSYS customers typically use 32 to 128 processors for fluids, while more complex mechanical simulations will use 16 to 32 cores.
In an effort to provide faster performance during solution, various ANSYS products support offloading key solver computations onto graphics cards to accelerate those computations.
ANSYS Mechanical users can accelerate both their shared-memory and distributed-memory parallel simulations using either single or multiple NVIDIA GPU cards or Intel Xeon Phi coprocessors on both Windows and Linux platforms. For example, ANSYS Mechanical users can get higher throughput on large models with an NVIDIA Tesla K40 in their workstations. Models as large as 12M degrees of freedom can be accelerated with a single K40 and provide 2X performance gains over just CPUs. Click here for a Tech Tip article on this subject. Similarly, adding one Intel Xeon Phi to 1-8 CPU core workstation leads to an average speed-up of 2X, particularly for the bulkier models.
ANSYS Fluent offers NVIDIA GPU support for a pressure-based coupled solver and for radiation heat transfer models. The current implementation accelerates the pressure-based coupled flow solver, speeding up the flow portion of CFD simulations. It is designed to run across multiple nodes with multiple GPUs in a cluster configuration, just like CPU systems. Please refer to the NVIDIA application note on Accelerating ANSYS Fluent 15.0 Using NVIDIA GPUs for more details.
ANSYS HFSS Transient can leverage multiple NVIDIA GPU cards with its hybrid solving technology for each excitation. Click here for a Tech Tip article demonstrating how this works . NVIDIA GPU cards are the officially supported platforms for use with ANSYS Mechanical. For more info, click here. For the circuit simulation engine provided with the ANSYS SI Option (i.e., Designer/Nexxim) the CUDA 2.0 platforms are supported for accelerating both Quickeye and IBIS-AMI analysis.
All HPC license products (HPC, HPC Packs, and HPC Workgroups) enable GPU-accelerated computing. Specifically, each GPU socket (or coprocessor) is treated the same as a single CPU core with respect to license requirements.
ANSYS HPC can tackle the largest simulations, solving problem with up to billions of elements for extreme high-fidelity simulation of complex geometries and small-scale physical phenomena. Learn how yacht designers and Formula One teams take advantage of this scalable architecture to consider problems that were considered extreme just a few years ago.
With ANSYS HPC, an IT expert or parallel processing guru isn't required. With a few simple menu picks, ANSYS software can be set up and launched to execute in parallel on a multi-core workstation or on a remote compute cluster. Everything needed to just install and start working, including MPI (message passing interface) tools, comes bundled with software from ANSYS.
Software from ANSYS is enabled to support load managers and job schedulers, including Platform LSF, PBS Professional and Microsoft Windows® HPC Server. Using GUI panels and/or customizable scripts, software from ANSYS can be integrated into your workload management environment with ease.
Courtesy BMW F1 Team.
ANSYS HPC provides parallel processing on a range of computing platforms. ANSYS HPC can be used on a multi-core workstation, substantially increasing throughput on single jobs. For workgroups or multiple job execution, clusters running Linux® or Microsoft Windows will provide scalable capacity and performance. ANSYS MPI-based parallel processing supports Gigabit Ethernet (GigE) or Infiniband® interconnects from a variety of vendors and the MPI tools required come fully bundled with the software. Software from ANSYS is also Intel® Cluster Ready (ICR) certified, ensuring smooth deployment and execution on ICR vendor solutions.
ANSYS makes a sustained investment in HPC software innovation, that culminates in the performance achieved when using ANSYS HPC solutions today. ANSYS is well recognized in the HPC industry and the company's HPC partners invest in ANSYS to ensure best-in-class performance on their latest technologies. Most importantly, HPC is an ongoing focus area for ANSYS, so the customer can be sure to gain value HPC technologies of the future. ANSYS is working now to enable simulations on 10,000 cores or more, with new innovative solver technologies, enhanced usability and scalability across the full simulation process. Choose ANSYS and be assured that your simulation provider has a roadmap into the HPC future.
ANSYS works with the leaders in HPC, including partners like Intel, AMD, Microsoft, HP, IBM, Dell, and other leading OEMs, resellers, technology vendors, and system integrators. Our technical collaboration with these partners ensures that ANSYS clients get the coordinated expert support needed at all phases of HPC deployment. From system specification to installation, tuning, troubleshooting and maintenance, ANSYS partners can help minimize risk and increase the productivity of HPC systems. Visit the partner solutions showcase and learn which partners are working closely with ANSYS to deliver HPC solutions tuned to customers' needs.