capture-kernel-trace

Official

Capture GPU kernel traces with ATT via rocprofv3.

AuthorROCm
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Capture GPU kernel ATT traces from a GPU environment (local or remote Docker), configure input.yaml for the target kernel, execute rocprofv3 -i input.yaml with source mapping enabled, and download the latest ui_output_agent_* data for analysis.

Core Features & Use Cases

  • Kernel discovery: identify available kernels on the target GPU environment to determine the tracing targets.
  • Input configuration: generate and tailor input.yaml with kernel_include_regex and tracing parameters.
  • Trace collection: run rocprofv3 -i input.yaml to collect Advanced Thread Trace data with DWARF mapping when enabled.
  • Trace retrieval: fetch and organize the latest ui_output_agent_* directory and associated trace artifacts for inspection.

Quick Start

Run the capture-kernel-trace skill with your test script and an optional kernel pattern to start collecting ATT traces.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: capture-kernel-trace
Download link: https://github.com/ROCm/FlyDSL/archive/main.zip#capture-kernel-trace

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.