README.md

Contribute to power profiling and model training

• Contribute to power profiling and model training
  • Requirements
  • Pre-step
  • Setup
    • Prepare cluster
    • From scratch (no target kubernetes cluster)
    • For managed cluster
    • Run benchmark and collect metrics
    • With manual execution
  • Clean up

Requirements

• git > 2.22
• hatch
• kubectl
• yq, jq
• power meter if available

Pre-step

• Fork and clone this repository and move to model_training folder

git clone
cd model_training

Setup

Prepare cluster

From scratch (no target kubernetes cluster)

Note: port 9090 and 5101 should not being used. It will be used in port-forward for prometheus and kind registry respectively

./script.sh prepare_cluster

The script will:

• create a kind cluster kind-for-training with registry at port 5101.
• deploy Prometheus.
• deploy Prometheus RBAC and node port to 30090 port on kind node which will be forwarded to 9090 port on the host.
• deploy service monitor for kepler and reload to Prometheus server

For managed cluster

Please confirm the following requirements:

• Kepler installation
• Prometheus installation
• Kepler metrics are exported to Promtheus server
• Prometheus server is available at http://localhost:9090. Otherwise, set environment PROM_SERVER.

Run benchmark and collect metrics

• Tekton Pipeline Instruction

With manual execution

In addition to the above approach, you can manually run your own benchmarks, then collect, train, and export the models by the entrypoint

Manual Metric Collection and Training with Entrypoint

Clean up

For kind-for-training cluster:

./script.sh cleanup