Platform and Hardware Requirements

Platforms, operating systems, and processors supported by nanoFluidX and recommended and required hardware.

Hardware

The hardware system that is intended to run nanoFluidX must contain a CUDA-enabled GPU. For Linux, the GPU must support Compute Capability 3.5 or higher. For Windows, the GPU must support Compute Capability 6.0 or higher. See https://developer.nvidia.com/cuda-gpus for more information.

NVIDIA Enterprise GPUs, data center, and RTX series are recommended GPU cards for HPC applications and nanoFluidX has been thoroughly tested on them.

Note: Each GPU generation and model has specific single precision and double precision performance.

NVIDIA GeForce line of GPU cards are also CUDA enabled and are capable of running nanoFluidX, however, NVIDIA GeForce Driver license agreement prohibits data center use. See https://www.nvidia.com/en-us/drivers/geforce-license for more information.

Hardware requirements are strongly dependent on the use case. For production level industrial cases, use the following minimum configuration:

  • The number of CPU cores should at least equal the number of GPU devices. Ideally, the number of CPU cores will slightly exceed the number of available GPU devices to ensure some computational overhead for system operations.
  • Minimum RAM per node is at least equal to the RAM of GPUs combined per node. Twice the combined GPU RAM size is recommended.
  • 3TB HDD space (long-term storage) or 500GB for operational drive. Common nanoFluidX output can vary from 20 to 400 GB, depending on the size of the case, requested outputs and frequency of the output.
  • High speed interconnect for multi-node systems, for example Infiniband.
Note:
CUDA Upgrade and NVIDIA Driver Support
In order to support the latest generation of NVIDIA GPUs, nanoFluidX 2024.1 was upgraded to CUDA 12 (currently CUDA 11) libraries. CUDA 12 removes support for older NVIDIA drivers, so a minimum driver version of 525.60.13 (Linux) / 528.33 (Windows) or above is required to run the solver.
Important: From version 2024.1 Kepler generation (K20/K40/K80) Tesla GPUs will no longer be supported.

Supported Platforms

Installing and running nanoFluidX on other supported Linux distributions may be possible, but we do not test, verify or warrant the functionality on these platforms.

Table 1. Officially Supported Platforms
OS Version
Windows 11
Windows 10
RHEL 8.4
Oracle Linux 8.4
Rocky Linux 8.4
SLES 15 SP3
Note:
RHEL 7 Support Deprecated
RHEL 7 and derivatives (eg CentOS) goes into "Extended Life Cycle Support" and only gets critical security fixes from 30 June 2024. RHEL7 was already removed from the officially supported platforms for HyperWorks since 2021.1, but it was still possible to run nanoFluidX on RHEL 7 in recent versions. With 2024.1 RHEL7 support is deprecated, and we will fully drop support in a future release.