Install NVIDIA GPU Drivers on Ubuntu 22.04

Based on Ubuntu 22.04

Prerequisites#

Hardware check

Verify that your system has a compatible NVIDIA GPU. If you do not know the exact GPU model in your system, run the following command:

Copy

1
lspci -v | egrep "NVIDIA"

This command should output one line for each installed NVIDIA GPU. Check whether your hardware is compatible on our Hardware Requirements page. Note that you can still install the CPU version of HEAVY.AI on machines without an NVIDIA GPU.

System update before installation

Upgrade the system and kernel, then reboot if needed.

Copy

1
sudo apt update

Copy

1
sudo apt upgrade -y
2
sudo reboot

Install kernel headers#

Install kernel headers and development packages.

Copy

1
sudo apt install linux-headers-$(uname -r)

Install additional packages.

Copy

1
sudo apt install pciutils

Install Vulkan libraries#

The HEAVY.AI rendering engine (available in the Enterprise Edition) requires drivers that support Vulkan and the Vulkan libraries. Without these components, the database itself might fail to start.

Use apt to install the Vulkan libraries and their dependencies.

Copy

1
sudo apt install libvulkan1

For more information on Vulkan troubleshooting, see the Vulkan Renderer section.

Install NVIDIA drivers#

Run this command to get a list of available driver versions:

1
apt list nvidia-driver-*

Use apt to install the desired driver version.

Copy

1
sudo apt install nvidia-driver-<version>

Post-install steps for NVIDIA drivers#

Reboot the system to ensure the new driver version is loaded.

Copy

1
sudo reboot

Verify that NVIDIA drivers are installed correctly#

Run nvidia-smi to verify that your drivers are installed correctly and that the GPUs in your environment are detected. Depending on your environment, you should see output similar to the following, confirming that your NVIDIA GPU and drivers are present.

Install Vulkan libraries#

The HEAVY.AI rendering engine requires drivers that support Vulkan and the Vulkan libraries. Without these components, the database itself cannot even start unless the backend renderer is disabled.

Use apt to install the Vulkan libraries and their dependencies.

Copy

1
sudo apt install libvulkan1

For more information on Vulkan troubleshooting, see the Vulkan Renderer section.

Install CUDA Toolkit#

If you installed the NVIDIA drivers using Option 1 above, the CUDA Toolkit is already installed; you can skip directly to the verification steps below.

Install the NVIDIA public repository GPG key.

Copy

1
distribution=$(. /etc/os-release;echo $ID$VERSION_ID | sed -e 's/\.//g')
2
sudo apt-key adv --fetch-keys \
3
https://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64/3bf863cc.pub

Add the repository.

Copy

1
echo "deb http://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64 /" \
2
| sudo tee /etc/apt/sources.list.d/cuda.list
3
apt update

List available CUDA Toolkit versions.

1
root@VM-8-208-ubuntu:~# apt list cuda-toolkit-* | grep -v config
2

3
WARNING: apt does not have a stable CLI interface. Use with caution in scripts.
4

5
Listing...
6
cuda-toolkit-11-7/unknown 11.7.1-1 amd64
7
cuda-toolkit-11-8/unknown 11.8.0-1 amd64
8
cuda-toolkit-12-0/unknown 12.0.1-1 amd64
9
cuda-toolkit-12-1/unknown 12.1.1-1 amd64
10
cuda-toolkit-12-2/unknown 12.2.2-1 amd64
11
cuda-toolkit-12-3/unknown 12.3.2-1 amd64
12
cuda-toolkit-12-4/unknown 12.4.1-1 amd64
13
cuda-toolkit-12-5/unknown 12.5.1-1 amd64
14
cuda-toolkit-12-6/unknown 12.6.3-1 amd64
15
cuda-toolkit-12/unknown 12.6.3-1 amd64

Use apt to install the CUDA Toolkit.

Copy

1
sudo apt install cuda-toolkit-<version>

Verification#

Check that everything is working properly and that the toolkit has been installed.

Copy

1
/usr/local/cuda/bin/nvcc --version
2

3
nvcc: NVIDIA (R) Cuda compiler driver
4
Copyright (c) 2005-2020 NVIDIA Corporation
5
Built on Mon_Nov_30_19:08:53_PST_2020
6
Cuda compilation tools, release 11.2, V11.2.67
7
Build cuda_11.2.r11.2/compiler.29373293_0

1
root@VM-8-208-ubuntu:~# echo 'export PATH=/usr/local/cuda/bin:$PATH' >> ~/.bashrc
2
root@VM-8-208-ubuntu:~# echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
3
root@VM-8-208-ubuntu:~# source ~/.bashrc
4
root@VM-8-208-ubuntu:~# nvcc --version
5
nvcc: NVIDIA (R) Cuda compiler driver
6
Copyright (c) 2005-2024 NVIDIA Corporation
7
Built on Tue_Oct_29_23:50:19_PDT_2024
8
Cuda compilation tools, release 12.6, V12.6.85
9
Build cuda_12.6.r12.6/compiler.35059454_0

Install Clang#

If you use advanced features such as C++ user-defined functions and/or user-defined table functions to extend database functionality, you must install Clang. Use apt to install Clang and its LLVM dependencies.

Copy

1
sudo apt install clang

Verification#

Verify that the software is installed and available on the execution path.

Copy

1
root@VM-8-208-ubuntu:~# clang --version
2
Ubuntu clang version 14.0.0-1ubuntu1.1
3
Target: x86_64-pc-linux-gnu
4
Thread model: posix
5
InstalledDir: /usr/bin

Install Docker#

1
apt install curl vim wget gnupg dpkg apt-transport-https lsb-release ca-certificates
2

3
curl -sS https://download.docker.com/linux/debian/gpg | gpg --dearmor > /usr/share/keyrings/docker-ce.gpg
4

5
echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-ce.gpg] https://mirrors.tuna.tsinghua.edu.cn/docker-ce/linux/ubuntu $(lsb_release -sc) stable" > /etc/apt/sources.list.d/docker.list
6

7
apt update && apt install docker-ce docker-ce-cli containerd.io docker-compose-plugin

1
root@catcat:~# cat /etc/docker/daemon.json
2
{
3
    "log-driver": "json-file",
4
    "log-opts": {
5
        "max-size": "20m",
6
        "max-file": "3"
7
    },
8
    "experimental":true,
9
    "data-root": "/root/docker_data",
10
    "registry-mirrors": ["https://docker.1ms.run" ]
11
}
12
systemctl restart docker

Install GPU-Docker components#

1
 Install gpu-docekr
2

3
apt-get install -y nvidia-docker2
4
nvidia-ctk runtime configure --runtime=docker
5

6
This modifies the daemon.json file to add the container runtime
7
root@VM-8-208-ubuntu:~# nvidia-ctk runtime configure --runtime=docker
8
INFO[0000] Loading config from /etc/docker/daemon.json
9
INFO[0000] Wrote updated config to /etc/docker/daemon.json
10
INFO[0000] It is recommended that docker daemon be restarted.

1
root@VM-8-208-ubuntu:~# cat /etc/docker/daemon.json
2
{
3
    "data-root": "/root/docker_data",
4
    "experimental": true,
5
    "log-driver": "json-file",
6
    "log-opts": {
7
        "max-file": "3",
8
        "max-size": "20m"
9
    },
10
    "registry-mirrors": [
11
        "https://docker.1ms.run"
12
    ],
13
    "runtimes": {
14
        "nvidia": {
15
            "args": [],
16
            "path": "nvidia-container-runtime"
17
        }
18
    }
19
}

Verify the installation#

1
root@VM-8-208-ubuntu:~# systemctl restart docker
2
root@VM-8-208-ubuntu:~# docker run --rm -it --gpus all ubuntu:22.04 /bin/bash
3
root@cad0abb4936b:/# nvidia-smi
4
Tue Aug  1 00:57:29 2023
5
+-----------------------------------------------------------------------------+
6
| NVIDIA-SMI 520.61.05    Driver Version: 520.61.05    CUDA Version: 11.8     |
7
|-------------------------------+----------------------+----------------------|
8
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
9
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
10
|                               |                      |               MIG M. |
11
|===============================+======================+======================|
12
|   0  NVIDIA GeForce ...  Off  | 00000000:01:00.0  On |                  N/A |
13
|  0%   51C    P8    21W / 160W |    528MiB /  6144MiB |     29%      Default |
14
|                               |                      |                  N/A |
15
+-------------------------------+----------------------+----------------------+
16

17
+-----------------------------------------------------------------------------+
18
| Processes:                                                                  |
19
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
20
|        ID   ID                                                   Usage      |
21
|=============================================================================|
22
+-----------------------------------------------------------------------------+
23
root@cad0abb4936b:/#

1
root@l4d2-forward:~/TeamSpeak# docker  logs teamspeak-teamspeak-1
2
2024-12-12 05:48:10.030543|INFO    |ServerLibPriv |   |TeamSpeak 3 Server 3.13.7 (2022-06-20 12:21:53)
3
2024-12-12 05:48:10.030628|INFO    |ServerLibPriv |   |SystemInformation: Linux 5.10.0-30-amd64 #1 SMP Debian 5.10.218-1 (2024-06-01) x86_64 Binary: 64bit
4
2024-12-12 05:48:10.030651|INFO    |ServerLibPriv |   |Using hardware aes
5
2024-12-12 05:48:10.030986|INFO    |DatabaseQuery |   |dbPlugin name:    MariaDB plugin, version 3, (c)TeamSpeak Systems GmbH
6
2024-12-12 05:48:10.031040|INFO    |DatabaseQuery |   |dbPlugin version: 2
7
2024-12-12 05:48:18.453775|INFO    |SQL           |   |db_CreateTables() tables created
8

9
------------------------------------------------------------------
10
                      I M P O R T A N T
11
------------------------------------------------------------------
12
               Server Query Admin Account created
13
         loginname= "serveradmin", password= "zKqn6Y0j"
14
         apikey= "BAAUOSo-F8ADMyaE9BqtVukv89P60wlNvFykSGq"
15
------------------------------------------------------------------
16

17
2024-12-12 05:48:18.515954|WARNING |Accounting    |   |Unable to open licensekey.dat, falling back to limited functionality
18
2024-12-12 05:48:18.524291|INFO    |Accounting    |   |Licensing Information
19
2024-12-12 05:48:18.524324|INFO    |Accounting    |   |licensed to       : Anonymous
20
2024-12-12 05:48:18.524344|INFO    |Accounting    |   |type              : No License
21
2024-12-12 05:48:18.524363|INFO    |Accounting    |   |starting date     : Tue Feb  1 00:00:00 2022
22
2024-12-12 05:48:18.524380|INFO    |Accounting    |   |ending date       : Thu Jul  1 00:00:00 2027
23
2024-12-12 05:48:18.524392|INFO    |Accounting    |   |max virtualservers: 1
24
2024-12-12 05:48:18.524403|INFO    |Accounting    |   |max slots         : 32
25
2024-12-12 05:48:19.071662|INFO    |              |   |Puzzle precompute time: 530
26
2024-12-12 05:48:19.072464|INFO    |FileManager   |   |listening on 0.0.0.0:30033, [::]:30033
27
2024-12-12 05:48:19.077083|INFO    |VirtualSvrMgr |   |executing monthly interval
28
2024-12-12 05:48:19.077549|INFO    |VirtualSvrMgr |   |reset virtualserver traffic statistics
29
2024-12-12 05:48:19.084343|INFO    |Query         |   |Using a query thread pool size of 2
30
2024-12-12 05:48:19.130955|INFO    |VirtualServerBase|1  |listening on 0.0.0.0:9987, [::]:9987
31
2024-12-12 05:48:19.135106|WARNING |VirtualServer |1  |--------------------------------------------------------
32
2024-12-12 05:48:19.135139|WARNING |VirtualServer |1  |ServerAdmin privilege key created, please use the line below
33
2024-12-12 05:48:19.135158|WARNING |VirtualServer |1  |token=BK2xL7Bm4Q2NEq4hsYYUJm4N1pVFuyZn6s33Jr6x
34
2024-12-12 05:48:19.135173|WARNING |VirtualServer |1  |--------------------------------------------------------
35

36
------------------------------------------------------------------
37
                      I M P O R T A N T
38
------------------------------------------------------------------
39
      ServerAdmin privilege key created, please use it to gain
40
      serveradmin rights for your virtualserver. please
41
      also check the doc/privilegekey_guide.txt for details.
42

43
       token=BK2xL7Bm4Q2NEq4hsYYUJm4N1pVFuyZn6s33Jr6x
44
------------------------------------------------------------------
45

46
2024-12-12 05:48:19.135520|INFO    |Query         |   |listening for query on 0.0.0.0:10011, [::]:10011
47
2024-12-12 05:48:19.135711|INFO    |CIDRManager   |   |updated query_ip_allowlist ips: 127.0.0.1/32, ::1/128,
48
root@l4d2-forward:~/TeamSpeak#