job-report

Generate HPC job summary reports from Omnistat GPU telemetry.

26|6|Updated Oct 9, 2023
One-click install
npx skills add https://github.com/ROCm/omnistat --skill job-report
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: job-report
Source: https://github.com/ROCm/omnistat/tree/main/skills/job-report
Command: npx skills add https://github.com/ROCm/omnistat --skill job-report

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires omnistat, python, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a fast and easy way to summarize the performance and resource utilization of HPC jobs, offering a high-level overview without in-depth analysis.

Core Features & Use Cases

  • Single-shot Report Generation: Generate a comprehensive report for an HPC job with a single command.
  • Job Summary: Capture essential metrics such as GPU utilization, power, temperature, and frequency.
  • Use Case: When you need a quick snapshot of a job's performance to identify trends or anomalies without the need for detailed analysis.

Quick Start

To generate a report for an HPC job with ID "JOB123", use the following command:

run omnistat-inspect job JOB123 report > report.json

Frequently Asked Questions about job-report

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate an HPC job performance summary using Omnistat?

To generate an HPC job performance summary with Omnistat, execute `omnistat-inspect job <JOB_ID> report` to output a JSON file containing GPU telemetry, utilization, power, and temperature metrics.

What is included in an HPC job summary report for resource utilization?

An HPC job summary report includes high-level resource utilization metrics such as GPU utilization, power consumption, temperature, and frequency to help identify performance trends without in-depth analysis.

Do I need a specific database to collect GPU telemetry for HPC jobs?

You need access to a time-series database (TSDB) or CSV data source to collect GPU telemetry. The reporting process also requires Python and the Omnistat package to query and summarize the HPC job data.

Can I use this approach for detailed HPC job performance analysis?

This approach is designed for quick performance overviews and identifying high-level trends rather than detailed analysis. It provides a fast snapshot of resource utilization but lacks in-depth analytical capabilities.

What's the best way to quickly identify anomalies in HPC job performance?

The best way to identify anomalies quickly is generating a single-shot summary report to capture essential GPU metrics like utilization and power, providing a fast snapshot of job performance without needing deep analysis.