v0.2.0-rc.1

• Fix nvidia/cuda version to 10.0-base
• Update nfd.yaml to v0.6.0 following example on official project page
• Update getArchFamily() to return newer microarchitecture information
• Add support for the 'none', 'single' and 'mixed' MIG strategies
• Add example deamonset deployments for NFD
• Add example deamonset deployments for different MIG strategies
• Add support for deploying gpu-feature-discovery as a helm chart
• Add support for deploying node-feature-discovery as a dependent subchart

v0.1.0

• Add label for GPU count

v0.0.0

This is the first beta release of NVIDIA GPU Feature Discovery

We may break the API in future releases. However, we will setup a deprecation policy.

• The label format changed from nvidia-* to nvidia.com/*

• All the labels are handled:

  • nvidia.com/cuda.runtime.major
  • nvidia.com/cuda.runtime.minor
  • nvidia.com/cuda.driver.major
  • nvidia.com/cuda.driver.minor
  • nvidia.com/cuda.driver.rev
  • nvidia.com/gpu.family
  • nvidia.com/gpu.machine
  • nvidia.com/gpu.product
  • nvidia.com/gpu.memory
  • nvidia.com/gpu.compute.major
  • nvidia.com/gpu.compute.minor
  • nvidia.com/gfd.timestamp

• We are using the require module for the tests

1.0.0-alpha

This is the first release of NVIDIA GPU Feature Discovery

• We are using the Node Feature Discovery to label nodes
• We are handling a simple set of labels:
  • nvidia-driver-version
  • nvidia-model
  • nvidia-memory