The Rocky Solver SDK is supported for the following operating systems:

• Centos 7 Linux 64-bit
• Windows 10 64-bit

This manual assumes that your system is one of the above and Rocky can perform a simulation successfully in this system.

Note: Refer to the Rocky User Manual or search on the Rocky support channels if you were unable to run Rocky.

The Rocky Solver SDK package for both supported operating systems can be downloaded from Ansys Customer Portal.

It is assumed that users of the Rocky Solver SDK have basic knowledge of the C++ and Python programming languages.

Prerequisites for CentOS 7 Linux

Unzip the Rocky Solver SDK for Linux in a folder of your choice. Set the environment variable RockySDK_DIR with the location of the unzipped sdk subdirectory.

The version of the Rocky Solver SDK is identified in the name of its zip file. It is paramount that you install a Rocky Solver SDK of the same version as the Rocky application upon which your module is designed to run.

Note: The version of Rocky can be discovered by choosing Help → About in the application menus.

Install the CUDA toolkit 11.7.0 for Linux x86_64, CentOS 7. Set the environment variable CUDA_PATH_V11_7 with the location of the installed CUDA toolkit. Example: CUDA_PATH_V11_7=/usr/local/cuda-11.7/

Build tools - Linux

Install Miniconda for Python 3.8 Linux 64-bit, then, activate a conda prompt and install the following packages:

• make
• cmake
• "gxx_linux-64>=11,<12"

The conda prompt now contains a minimum set of tools required to build Rocky modules.

Prerequisites for Windows 10

Unzip the Rocky SDK for Windows in a folder of your choice. Set the environment variable RockySDK_DIR with the location of the unzipped sdk subdirectory.

The version of the Rocky SDK is identified in the name of its zip file. It is paramount that you install a Rocky SDK of the same version as the Rocky application upon which your module is designed to run.

Note: The version of Rocky can be discovered by choosing Help → About in the application menus.

Install CUDA toolkit 11.7.0 for Windows 10. Set the environment variable CUDA_PATH_V11_7 with the location of the CUDA toolkit. Example: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7

Build tools - Windows

Install Visual Studio Build Tools 2019, version 16.8.4.

Be sure to install the Desktop development with C++ workload with at least the following components:

• VC++ 2019 version latest tools
• Windows 10 SDK
• Visual C++ tools for CMake