GPU Checklist
In order to use GPU processing in Huygens, you will need the following:
- A graphics card with a NVIDIA GPU that was released after 2013/14 (Compute capability of 5.0 or higher);
- A computer with either Windows or Linux as operating system (Unfortunately, macOS does not support NVIDIA GPUs);
- An up-to-date NVIDIA driver for your graphics card;
- The NVIDIA CUDA toolkit software installed on your system;
- A Huygens licence that supports your graphics card.
More details will be provided below.
Graphics cards supported by Huygens Huygens only supports graphics cards with a NVIDIA graphics processing unit (GPU). Furthermore, the GPU should have a compute capability of 5.0 or higher. All NVIDIA GPUs released in 2015 or later should satisfy this. You can check the compute capability of each NVIDIA GPU on this page.
Installing and updating drivers for NVIDIA GPUs In order for your graphics card to function correctly, it needs an appropriate driver. When this driver is missing, or it is too old, Huygens might not be able to use your graphics card. If no drivers are installed yet, you can download them from NVIDIA's website. Note that administrator privileges are required to install graphics drivers.
If you have already installed drivers and you want to update them, you can do so as follows (administrator privileges required):
- On Windows:
You can simply download the latest version of your driver from NVIDIA's website, the installer will take care of the rest.
- Find an appropriate driver on NVIDIA's website.
- Follow the steps on the website to download the installer.
- Execute the installer and follow the steps therein (This requires administrator privileges).
- Restart your computer.
- On Linux:
- Open "Software & Updates";
- Select the "Additional Drivers" tab;
- Select the latest NVIDIA driver in the list;
- Press "Apply Changes" and wait for the installation to complete;
- Restart your computer.
Installing the NVIDIA CUDA Toolkit The NVIDIA CUDA Toolkit is a software package that enables your GPU to be used for high-performance computing. It needs to be installed if you want to use GPU processing in Huygens. You can download the latest version of the CUDA toolkit here. Newer CUDA versions might drop support for older GPUs, in this case it might be needed to download an older version of CUDA. You can find which CUDA version supports which compute capability here and find the compute capability of your graphics card here.
Huygens needs at least CUDA version 11.0. The latest CUDA version that is confirmed to work with Huygens is 12.2 (last updated: 2023-11). Usually newer versions work as well.
GPU licensing Most standard graphics cards used in laptops and desktops are supported by default. For high performance GPU cards (typically used in dedicated workstations, servers and super computers), we have different tiers of GPU options. You can find out which license you need for your GPU card on this page. To see whether your current license fits the specifications of your GPU card, open the Huygens License manager (via the Help menu). Under the section "Technical Information" you see whether you are using a node-locked or Everywhere license and what graphics cards are supported (small, medium, large, extreme).
Troubleshooting
Checking the status of GPU acceleration The GPU status window is available in the toolbar by default and accessible in through the main menu under Help. In this window, you can check the status of each graphics card. There is also a tab with troubleshooting info as well as a detailed report on the status of Huygens GPU processing.
Checking if drivers have been installed You can test if you have a working graphics driver by running the "nvidia-smi" command line tool.
On Windows, it is typically installed in C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe. You can run it like this: 1) Open a cmd.exe command prompt by typing: <windows-key> + r and then entering cmd in the window that appears. 2) Change the current directory to the correct folder by typing:
cd "\Program Files\NVIDIA Corporation\NVSMI"
3) Run nvidia-smi by typing:
nvidia-smi
An alternative location where nvidia-smi.exe could be installed on Windows is in C:\Windows\System32\nvidia-smi.exe.
On Linux, this is typically installed in the search path for executables, so you can run it by opening a command line terminal and typingnvidia-smi
If succesful, it should report the graphics driver version in the top left, the maximum supported CUDA Toolkit version (which is not necessarily the same thing as the currently installed CUDA Toolkit version) in the top right, and the status of all detected graphics cards. If there are errors, then the installation of your graphics driver plus CUDA Toolkit combination is not working properly. After installation of new versions, a reboot is often necessary to activate the new graphics driver.
Note that nvidia-smi does not tell you if CUDA has been installed. For this, use the steps below.
Checking if CUDA has been installed You can test if CUDA has been installed via the command line:
On Windows: 1) Open a cmd.exe command prompt by typing: <windows-key> + r and then entering cmd in the window that appears.
2) Check if CUDA is installed by typingnvcc --version
nvcc --version
When CUDA has been correctly installed, you should see something like this:
nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2016 NVIDIA Corporation Built on Mon_Jan__9_17:32:33_CST_2017 Cuda compilation tools, release 8.0, V8.0.60
When you get a message that nvcc does not exist, then CUDA has likely not been installed, and you will have to follow the steps under Installing the NVIDIA CUDA Toolkit above to install it.
Contact support In case the steps above are not sufficient to solve your problem, please contact our support team. When doing so, please provide the following such that we can help you as quickly as possible:
- A screenshot of the "Full Report" tab in the GPU status window.
- An execution log file. See here for where to find this log.