System requirements for Ansys
-Ansys hardware requirements-
The minimum requirements for Ansys Discovery software are as follows:
– 64-bit Intel or AMD system with Windows 10.
– 8 GB of RAM
– A dedicated graphics card with the latest drivers and at least 1GB of video RAM that can support OpenGL 4.5 and DirectX 11 or higher. The use of integrated graphics (eg Intel HD/IRIS) is not recommended and is not supported during the analysis phase in Discovery. See below for special graphics requirements for ANSYS Discovery Live.
– button mouse
–Ansys Discovery Live and Explore Stage Graphics required
Ansys Discovery Live, or the exploration stage in Ansys Discovery, relies on the latest GPU technology to deliver its calculations and visual experience. To run the software, you need the following:
The Discovery Live Compatibility Tool is run after installation to check if your current graphics hardware is capable of supporting Ansys Discovery Live. If you do not have a graphics card with these specifications, the software will not run. Also, please make sure you have the latest driver for your graphics card available from NVIDIA Driver Downloads.
– Servers vs workstations
If you have or will have 32 or more parallel licenses (HPC Packs or Enterprise Licenses) in the future, and you have one or more users who need to submit their jobs using a higher number of cores than is available on their current workstations, we tend to We are recommending servers (or a cluster) and a supported Job Scheduler is required.
– Operating system
OS Platform Support | OS Platform Support – By Application
Ansys products are supported on 64-bit operating systems. Ansys Mechanical Ansys Fluent Most of our customers are successfully running Ansys software on Windows 10 workstations. For an overview of supported operating systems, see the links above.
The server operating system (Windows Server or Red Hat Linux/SUSE Enterprise Linux) is required under the following conditions:
– Processor
The latest 64-bit multi-core Intel Xeon and AMD processors with the highest clock speed and number of cores available are recommended. Hyper Threading does not improve the speed of simulations, always evaluate the number of physical cores for Ansys simulation. Always try to get the latest CPU architecture version, even if the clock speed or core count doesn’t seem to have improved. Today’s CPUs are almost twice as fast as CPUs from 3 years ago listed at the same clock speed.
Note that Windows 7/8/10 only supports a maximum of two physical CPUs. For more than two physical CPUs, a Windows Server or Linux operating system is required.
– memory
At least 16 GB of memory is recommended. It is best to have as much memory as is financially possible. The actual memory required for a particular problem depends on the mesh, active physical models, and domain complexity. As of 2016, 64GB of memory was sufficient for 90% of FEA and CFD projects completed by OEI engineers. EMAG products often require more memory and 100+ GB is recommended.
In terms of memory impact on performance, you either have enough or you don’t. If your operating system runs out of memory, it will use the hard disk as virtual memory, which will have a disastrous effect on system performance.
To understand why this is, it’s helpful to look at how the CPU works. CPUs have very little memory that they can access immediately. We register this. To access something that is not already in the register, the CPU must wait for the process of loading it into the register to complete before it can continue. There are different levels of memory in ascending size and descending performance that the CPU can access. Cache levels are located directly on the CPU itself and have different levels called L1, L2, etc… System memory or RAM is added modularly to the motherboard, like a hard disk. To compare the relative speed of these memory levels, we can use the metaphor of this excellent article on the subject:
This is not money to push pennies on system memory!
– Storage
At least 1TB is recommended for installing and using your Ansys software. An efficient approach to storage for Ansys might include: a smaller, faster drive (NVMe) for solving and a larger, slower drive (Mechanical, SSD) for storage. In Ansys Mechanical, you can specify a Solver Scratch Directory to ensure that solvers are automatically performed on high-performance drives, but stored in public memory.
The exact impact of storage on performance depends on how the I/O of a particular analysis type is constrained, but is uniformly better as model sizes get larger. If the type of analysis you expect is one of the following where I/O is typically a bottleneck in performance, strongly consider one of the following advanced storage recommendations:
– Types of input/output limit analysis:
There are several ways to increase storage performance:
NVMe: NVMe prices have recently come down and performance has improved. NVMe (Non-Volatile Memory Express) is the most recommended type of storage for high performance requirements in reading and writing data.
SSD drives: On a per gigabyte basis, they are significantly more expensive than mechanical hard drives, but can offer 2x faster read performance and 1x faster write performance. Make sure you have a modern operating system with TRIM support, otherwise the SSD’s write performance will degrade over time.
RAID0: While there are many different RAID configurations that vary in speed, redundancy, and efficient use of space, RAID0 is the only configuration that should be considered for performance. Redundancy should only be considered for individual “storage” drives or arrays, especially since RAID0 sacrifices performance the most (if any drive in a RAID0 array fails, all data is lost).
SSD + RAID0: Make sure TRIM is specifically supported for RAID0 with your chosen brand and operating system. SSD support on its own does not mean it supports TRIM in RAID0 arrays, which is a very recent development that often requires the latest operating system (like Windows 10) and drivers.
– Graphics card
Supported graphics cards
For large collections, it is recommended to use a graphics card to avoid display lag issues. You can find the list of valid graphics cards in the links above.
On-board graphics reduce the on-device pre- and post-processing experience. Additionally, newer, more graphics-intensive programs such as AIM and SpaceClaim will not work without a discrete graphics card.
– GPU
GPU computing resources | Supported GPU cards
In an effort to provide faster performance during solving, various Ansys products (Ansys Mechanical & Ansys Fluent) support offloading key solver calculations to graphics cards to accelerate these calculations. All HPC license products (HPC, HPC packages, and HPC workgroups) enable GPU-accelerated computing, and one GPU counts as one core. Note that not all graphics cards have CUDA support, this feature is reserved for high-end NVIDIA Tesla and Intel Phi cards. It is recommended to get a card with the largest amount of memory. Supported cards for GPU computing can be found in the document linked above:
Notes on GPU computing:
NVIDIA Tesla solutions from 2017 have more complete features. Not all analysis types are supported for GPU calculations.
GPU computing is currently well-suited to specific types of problems:
– HPC
Overview of HPC features:
To take full advantage of your computer’s hardware, make sure you have the appropriate HPC licenses. There are flexible HPC, HPC Pack and HPC Workgroup options. HPC licenses are on a per-core basis. A single GPU is licensed as a single core. HPC packages are added non-linearly to quickly access large numbers of cores.
An example of a desktop workstation
The following specification is an example of a mid-range desktop workstation used by a typical Ansys analyst. Please note that these systems were configured in 2020 and still perform well using Ansys 2020R1. Other server hardware is available for more demanding tasks.
Related links