hy-smi

Diagnose and configure Hygon GPU/DCU hardware via the hy-smi command-line tool.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/dongg622/china-ai-chip-skill --skill hy-smi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hy-smi
Source: https://github.com/dongg622/china-ai-chip-skill/tree/main/Hygon/hy-smi
Command: npx skills add https://github.com/dongg622/china-ai-chip-skill --skill hy-smi

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines the management and health monitoring of Hygon DCU hardware, enabling system administrators to quickly diagnose, check status, and troubleshoot device issues.

Core Features & Use Cases

  • Device health inspection: Perform comprehensive device health checks and error diagnostics to ensure optimal operation.
  • Performance tuning: Adjust power and clock settings to optimize GPU/DCU performance for intensive workloads.
  • Real-time monitoring: Continuously observe power usage, temperature, and utilization metrics during operation.
  • Fault troubleshooting: Detect and reset errors, analyze RAS logs, and diagnose hardware faults such as ECC errors or XGMI issues.
  • Device topology insight: Visualize and examine device interconnections to facilitate maintenance and configuration.

Quick Start

Use the hy-smi skill to perform a health check on your Hygon GPU/DCU system.

Frequently Asked Questions about hy-smi

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

FAQPage Schema
How do I check the health and status of my Hygon DCU hardware?

To check Hygon DCU hardware health, you can perform comprehensive device status checks and error diagnostics. This process inspects hardware operation metrics and pinpoints existing device faults to ensure optimal performance.

How do I troubleshoot ECC or XGMI faults on Hygon GPU devices?

Troubleshooting ECC or XGMI faults on Hygon GPU devices involves detecting and resetting errors alongside analyzing RAS logs. This fault diagnosis identifies specific hardware issues to restore proper device operation.

Can I monitor power usage and temperature metrics for DCU hardware in real-time?

Yes, you can continuously monitor DCU hardware metrics in real-time. This observes power usage, temperature, and utilization data during operation to track device behavior and maintain system stability.

What is the best way to optimize GPU performance for intensive workloads on Hygon devices?

The best way to optimize Hygon GPU performance is through performance tuning. This adjusts power and clock settings specifically for intensive workloads, maximizing hardware output and efficiency.

How can I visualize device topology and interconnections for Hygon DCU hardware?

You can visualize Hygon DCU device topology by examining device interconnections. This topology analysis maps hardware connections to facilitate maintenance planning and configuration management.

Do I need the hy-smi command-line tool to manage Hygon DCU devices?

Yes, you need the hy-smi command-line tool to manage Hygon DCU devices. It acts as the required interface to communicate directly with device firmware and drivers for hardware administration.