Install Nvidia driver, CUDA Toolkit and GPU-Burn without Docker

Overview

This instruction provides a step-by-step guide on selecting the correct NVIDIA driver and the corresponding CUDA Toolkit version, as well as how to install and run gpu-burn without using Docker.

Install Nvidia driver

First, check the avaible module versions

apt list | grep nvidia-driver

IMPORTANT: All RTX 50-series NVIDIA GPUs, at least until August 2025, require open source drivers!

And install the desired (always the latest) version driver

sudo apt install nvidia-driver-570=570.124.06-0ubuntu1

After the installation, verify whether it was successful and confirm the installed driver version

nvidia-smi

202508 Memo

Installing the newest driver is not always the best option, because NVIDIA driver updates and Ubuntu compatibility remain unstable, even though the 50-series GPUs have been on the market for about six months.

For example, in August 2025, after the release of driver 580.97, the 575-open and 570-open drivers cannot be installed on Ubuntu 24.04.3 LTS and show the following error:

Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

the following package have unmet dependencies : 
libnvidia-compute-575: depends libssl1.1(>=1.1.0) but it is not installable 

E : unable to correct problems, you have held broken package

Moreover, during the first week of release, installing 580 on Ubuntu 24.04 caused the system to freeze at the boot screen after reboot.

Therefore, I recommend using the 575 driver on Ubuntu 22.04.05 LTS until the 580 driver becomes stable.

Install Nvidia CUDA Toolkit

First, we need to install CUDA Toolkit.

1. Check CUDA GPU Compute Capability to choose proper CUDA version for your GPU Compute capability (CC) defines the hardware features and supported instructions for each NVIDIA GPU architecture. Find the compute capability for your GPU in the table . Example with the 50-series GPU: We need to install at least CUDA Toolkit version 12.0, otherwise subsequent tasks (e.g., running gpu-burn) may fail. More about CUDA version : When you use nvidia-smi to check the CUDA version supported by the driver, you may notice that it's different from the version of the installed CUDA toolkit. Because Ubuntu’s apt typically installs older toolkit versions, so your CUDA toolkit is often older than what nvidia-smi shows.

   That’s totally fine — CUDA drivers are backward compatible, so newer drivers can run code built with older toolkits.

2. Go to the CUDA Toolkit 12.0 Downloads and follow the installation instructions, make install CUDA toolkit on ubuntu 22.04 for example :

   wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
   sudo mv cuda-ubuntu2204.pin /etc/apt/preferences.d/cuda-repository-pin-600
   wget https://developer.download.nvidia.com/compute/cuda/12.0.0/local_installers/cuda-repo-ubuntu2204-12-0-local_12.0.0-525.60.13-1_amd64.deb
   sudo dpkg -i cuda-repo-ubuntu2204-12-0-local_12.0.0-525.60.13-1_amd64.deb
   sudo cp /var/cuda-repo-ubuntu2204-12-0-local/cuda-*-keyring.gpg /usr/share/keyrings/
   sudo apt-get update
   sudo apt-get -y install cuda

3. Install cuDNN (Optional but Recommended)

   For deep learning applications, install cuDNN:

   sudo apt install cudnn9-cuda-12

4. Verify Installation

   Check if CUDA is properly installed:

   nvcc --version
   nvidia-smi

Install gpu-burn and Run

After cloning the gpu-burn repository, simply run make and execute the gpu-burn .

git clone https://github.com/wilicc/gpu-burn.git
cd gpu-burn
make
./gpu_burn 10000

Result

You can see the detail information of the GPU real-time by :

watch -n 1 nvidia-smi

Every 1.0s: nvidia-smi                                                                                                                                                             neou@neou-Nuvo-10000-Series:~

Thu Mar 27 16:27:04 2025
+-------------------------------------------------------------------------------+
| NVIDIA-SMI 570.124.06      Driver Version: 570.124.06    CUDA Version: 12.8   |
|---------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M  | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap  |         Memory-Usage | GPU-Util  Compute M. |
|                                 |                      |               MIG M. |
|=================================+======================+======================|
| 0 NVIDIA GeForce RTX 5070 Ti Off| 00000000:02:00.0 Off |                  N/A |
|  71%   71C    P1    299W / 300W |  14088MiB / 16303MiB |    100%      Default |
|                                 |                      |                  N/A |
+---------------------------------+----------------------+----------------------+

+-------------------------------------------------------------------------------+
| Processes:                                                                    |
|  GPU   GI   CI        PID   Type   Process name                    GPU Memory |
|        ID   ID                                                     Usage      |
|===============================================================================|
|    0   N/A  N/A       2225     G   /usr/lib/xorg/Xorg                   4MiB  |
|    0   N/A  N/A       9513     C   ./gpu_burn                       14056MiB  |
+-------------------------------------------------------------------------------+