Getting Started with Graphcloud

• 1. Graphcloud overview
  • 1.1. Prerequisites
    • 1.1.1. Linux and Python commands
    • 1.1.2. Managing the IPU hardware resources
• 2. Accessing Graphcloud
  • 2.1. Your account
    • 2.1.1. Logging in
    • 2.1.2. Adding more users
    • 2.1.3. Extending access to Graphcloud
    • 2.1.4. Closing your account
  • 2.2. Configuration overview
    • 2.2.1. Local storage
  • 2.3. Software and datasets
  • 2.4. Making efficient use of local storage
    • 2.4.1. Shared workspace
    • 2.4.2. Installing the Graphcore examples
    • 2.4.3. Temporary storage
• 3. Software setup
  • 3.1. Setting up the SDK environment
  • 3.2. Setting up PyTorch for the IPU
  • 3.3. Setting up TensorFlow for the IPU
    • 3.3.1. IPU TensorFlow Addons
    • 3.3.2. Next steps
  • 3.4. Verifying the hardware configuration
• 4. Running a program on the IPU
  • 4.1. Hardware and software monitoring
    • 4.1.1. Command line tools
    • 4.1.2. Trace output
    • 4.1.3. Execution profiling
• 5. Troubleshooting
• 6. Next steps
• 7. Trademarks & copyright