Skip to main content

Installing NVIDIA Drivers (SXM)

Introduction​

NVIDIA SXM GPUs are specialized for high-performance computing and deep learning applications, offering superior performance and efficiency compared to standard PCIe GPUs.

Installation Steps for NVIDIA CUDA 13.0 (NVIDIA driver 580) on H100/H200 SXM​

Ubuntu 24.04​

#!/bin/bash 

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb && 
sudo dpkg -i cuda-keyring_1.1-1_all.deb && 
sudo apt-get update && 
sudo apt-get -y install cuda-toolkit-13-0 &&
sudo apt-get install -y nvidia-driver-580-open && 
sudo apt-get install -y cuda-drivers-580 && 
echo "options nvidia NVreg_NvLinkDisable=1" | sudo tee /etc/modprobe.d/disable-nvlink.conf &&
sudo update-initramfs -u && sudo reboot

Ubuntu 22.04​

#!/bin/bash 

sudo mkdir -p /etc/needrestart/conf.d && 
echo "\$nrconf{restart} = 'a';" | sudo tee /etc/needrestart/conf.d/99-no-prompt.conf && 
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb && 
sudo dpkg -i cuda-keyring_1.1-1_all.deb && 
sudo apt-get update && 
sudo apt-get -y install cuda-toolkit-13-0 &&
sudo apt-get install -y nvidia-driver-580-open && 
sudo apt-get install -y cuda-drivers-580 && 
echo "options nvidia NVreg_NvLinkDisable=1" | sudo tee /etc/modprobe.d/disable-nvlink.conf &&
sudo update-initramfs -u && sudo reboot

Ubuntu 20.04​

NVIDIA doesn't ship the CUDA 13.0 / driver 580 stack for Focal, so Ubuntu 20.04 tops out at CUDA 12.9 and driver 575.

#!/bin/bash 

sudo mkdir -p /etc/needrestart/conf.d && 
echo "\$nrconf{restart} = 'a';" | sudo tee /etc/needrestart/conf.d/99-no-prompt.conf && 
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-keyring_1.1-1_all.deb && 
sudo dpkg -i cuda-keyring_1.1-1_all.deb && 
sudo apt-get update && 
sudo apt-get -y install cuda-toolkit-12-9 &&
sudo apt-get install -y nvidia-driver-575-open && 
sudo apt-get install -y cuda-drivers-575 && 
echo "options nvidia NVreg_NvLinkDisable=1" | sudo tee /etc/modprobe.d/disable-nvlink.conf &&
sudo update-initramfs -u && sudo reboot

Verification​

After reboot, reconnect to the VM, and verify the driver is installed and its version.

nvidia-smi

Expected output on Ubuntu 24.04 and 22.04. On Ubuntu 20.04 the header line shows NVIDIA-SMI 575.57.08, Driver Version: 575.57.08 and CUDA Version: 12.9 instead; the rest of the layout is the same.

+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.173.02             Driver Version: 580.173.02     CUDA Version: 13.0     |
|-----------------------------------------+------------------------+----------------------+
| 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 H100 80GB HBM3          Off |   00000000:09:00.0 Off |                    0 |
| N/A   36C    P0             69W /  700W |       1MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

Verify NVlink is disabled.

nvidia-smi -q | grep -A5 Fabric

Expected output:

    Fabric
        State                             : N/A
        Status                            : N/A
        CliqueId                          : N/A
        ClusterUUID                       : N/A
        Health

This needs to be N/A, NOT In Progress.

Contents