V-IPU User Guide

• 1. Introduction
  • 1.1. Scope of the document
  • 1.2. Structure of the document
• 2. Concepts and architecture
  • 2.1. Architecture
• 3. Getting started
  • 3.1. Installing the V-IPU client
  • 3.2. V-IPU configuration
  • 3.3. Creating partitions
• 4. Partitions
  • 4.1. Overview
  • 4.2. Creating a reconfigurable partition
    • 4.2.1. Limitations
  • 4.3. Creating a preconfigured partition
  • 4.4. IPU selection
  • 4.5. Displaying partition information
    • 4.5.1. Partition state
    • 4.5.2. Provisioning state
  • 4.6. IPUoF environment variables
  • 4.7. IPUoF configuration files
    • 4.7.1. Creating an IPUoF configuration file
    • 4.7.2. Sharing access to devices
  • 4.8. Removing a partition
  • 4.9. Routing configuration
    • 4.9.1. Intra-GCD sync configuration
  • 4.10. Multi-GCD partitions
    • 4.10.1. Multi-host IPUoF configuration
    • 4.10.2. Intra-GCD sync configuration
    • 4.10.3. Inter-GCD sync configuration
    • 4.10.4. Crossing IPU-Link domains
• 5. Integration with Slurm
  • 5.1. Configuring Slurm to use V-IPU select plugin
  • 5.2. Configuration parameters
  • 5.3. The V-IPU job submission plugin
  • 5.4. An example Slurm Controller configuration
  • 5.5. Job submission and parameters
    • 5.5.1. Job script examples
• 6. Integration with Kubernetes
  • 6.1. Components and design
  • 6.2. Package contents
  • 6.3. Deploying the software
    • 6.3.1. Prerequisites
    • 6.3.2. Installation
    • 6.3.3. Uninstall
    • 6.3.4. Upgrading the Helm Chart
  • 6.4. Configurations
  • 6.5. Creating an IPUJob
    • 6.5.1. Interactive mode
    • 6.5.2. Mounting data volumes for an IPUJob
    • 6.5.3. Automatic restarts
    • 6.5.4. Clean up Kubernetes resources and IPU partitions
  • 6.6. Debugging problems
    • 6.6.1. How does the IPU Operator work?
    • 6.6.2. Debugging
  • 6.7. IPU usage statistics
  • 6.8. Operator Metrics
  • 6.9. Known limitations
• 7. Command line reference
  • 7.1. Resource types
  • 7.2. Global options
    • 7.2.1. Using a configuration file
    • 7.2.2. Using environment variables
  • 7.3. Create partition
    • 7.3.1. Syntax
    • 7.3.2. Examples
    • 7.3.3. Options
  • 7.4. Get partition
    • 7.4.1. Syntax
    • 7.4.2. Examples
    • 7.4.3. Options
  • 7.5. List allocation
    • 7.5.1. Syntax
    • 7.5.2. Examples
    • 7.5.3. Options
  • 7.6. List IPUs
    • 7.6.1. Supported options
  • 7.7. List partition
    • 7.7.1. Syntax
    • 7.7.2. Examples
    • 7.7.3. Options
  • 7.8. Remove a partition
    • 7.8.1. Syntax
    • 7.8.2. Examples
    • 7.8.3. Options
  • 7.9. Reset partition
    • 7.9.1. Syntax
    • 7.9.2. Examples
    • 7.9.3. Options
  • 7.10. Bash completion
    • 7.10.1. Syntax
    • 7.10.2. Examples
    • 7.10.3. Options
  • 7.11. Zsh completion
    • 7.11.1. Syntax
    • 7.11.2. Examples
    • 7.11.3. Options
• 8. Revision history
• 9. Trademarks & copyright