DGX-1 User Guide
Documentation for administrators that explains how to install and configure the NVIDIA DGX-1 Deep Learning System, including how to run applications and manage the system through the NVIDIA Cloud Portal.
Table of Contents
- 1. Introduction to the NVIDIA DGX-1 Deep Learning System
- 2. Installation and Setup
- 2.1. Registering Your DGX-1
- 2.2. Choosing a Setup Location / Site Preparation
- 2.3. Unpacking the DGX-1
- 2.4. What's In the Box
- 2.5. Installing the DGX-1 Into a Rack
- 2.6. Attaching the Bezel
- 2.7. Connecting the Power Cables
- 2.8. Connecting the Network Cables
- 2.9. Setting Up the DGX-1
- 2.10. Post Setup Instructions for DGX OS Server Software Version 2.x and Earlier
- 2.11. Updating the DGX-1 Software
- 2.12. Managing CPU Mitigations
- 3. Preparing for Using Docker Containers
- 3.1. Installing Docker and the Docker Engine Utility for NVIDIA GPUs on DGX OS Server Software 2.x or Earlier
- 3.2. Configuring Docker IP Addresses
- 3.3. Letting Users Issue Docker Commands
- 3.4. Enabling GPU Support for NGC Containers
- 3.5. Configuring a System Proxy
- 3.6. Configuring NFS Mount and Cache
- 4. Configuring and Managing the DGX-1
- 4.1. Using the BMC
- 4.2. Configuring a Static IP Address for the BMC
- 4.3. Configuring Static IP Addresses for the Network Ports
- 4.4. Obtaining MAC Addresses
- 4.5. Resetting GPUs in the DGX-1
- 4.6. Changing the Mellanox Card Port Type
- 4.7. Enabling USB 3.0
- 5. Security
- 6. Maintaining and Servicing the NVIDIA DGX-1
- 6.1. Problem Resolution and Customer Care
- 6.2. Restoring the DGX-1 Software Image
- 6.3. Updating the System BIOS
- 6.4. Updating the BMC
- 6.5. Updating Component Firmware Using the Firmware Update Container
- 6.6. Replacing the System and Components
- 6.6.1. Replacing the System
- 6.6.2. Replacing an SSD
- 6.6.3. Recreating the Virtual Drives
- 6.6.4. Recreating the RAID 0 Array
- 6.6.5. Replacing the Power Supplies
- 6.6.6. Replacing the Fan Module
- 6.6.7. Replacing the Battery
- 6.6.8. Replacing the DIMMs
- 6.6.9. Installing/Replacing the 10GbE Mezzanine SPF+ NIC
- 6.6.10. Replacing the InfiniBand Cards
- 6.6.11. Setting Up the InfiniBand Cards
- 6.7. Secure Data Deletion of the SSDs
- 7. Installing Software on Air-Gapped NVIDIA DGX-1 Systems
- 8. Customer Support for the NVIDIA DGX-1
- 9. King Slide - AH61-500 Instructions
- 10. Safety
- 11. Compliance
- Notices