croq-dsl-helion

Validate environments, build, run, and profile Helion kernels with NVIDIA Nsight Compute.

Updated Apr 10, 2026
One-click install
npx skills add https://github.com/LancerLab/croqtile-tuner --skill croq-dsl-helion
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: croq-dsl-helion
Source: https://github.com/LancerLab/croqtile-tuner/tree/main/.claude/skills/croq-dsl-helion
Command: npx skills add https://github.com/LancerLab/croqtile-tuner --skill croq-dsl-helion

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill streamlines the tuning process for Helion kernels by providing detailed environment validation, build and run templates, profiling instructions, and safety checks, enabling efficient performance optimization.

Core Features & Use Cases

  • Environment Validation: Checks Helion version compatibility to ensure proper setup.
  • Build/Run Templates: Provides scripted templates for building and executing kernels, facilitating quick deployment.
  • Profiling Guidance: Guides users through using NVIDIA Nsight Compute (ncu) for performance analysis of Helion kernels.
  • Optimization Strategy: Offers configuration options and ideas for tuning kernel performance by adjusting parameters such as block sizes and warps, supporting iterative development.
  • Verification and Measurement: Ensures kernels produce correct results and measures throughput using standard timing procedures.
  • Baseline Setup: Suggests starting points with simple library calls for initial performance benchmarking.

Quick Start

Load the Helion DSL skill, validate your environment with the provided commands, and then follow the templates to build and profile your Helion kernels directly from your development environment.

Frequently Asked Questions about croq-dsl-helion

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

FAQPage Schema
How do I profile Helion kernels using NVIDIA Nsight Compute?

To profile Helion kernels, you can use guided instructions and scripted templates to execute NVIDIA Nsight Compute (ncu) for performance analysis. This structured workflow ensures effective profiling and safe measurement during kernel development.

What is the best way to tune GPU kernel performance in Helion?

The best way to tune GPU kernel performance in Helion is by adjusting parameters like block sizes and warps using provided configuration guidance. Iterative development and optimization strategies help systematically enhance kernel throughput.

How do I validate my environment before building Helion kernels?

To validate your environment for Helion kernels, run the provided commands to check Helion version compatibility. This validation ensures proper setup and prevents compatibility issues before you start building and running kernels.

Can I use standard Python library calls as a baseline for Helion kernel benchmarking?

Yes, you can use standard Python library calls as a baseline for Helion kernel benchmarking. The Skill suggests starting points with simple library calls to establish initial performance benchmarks before deeper optimization.

How do I verify correctness and measure throughput for Triton kernels?

To verify correctness and measure throughput for Triton kernels, use the provided build and run templates with standard timing procedures. This ensures kernels produce correct results while accurately measuring performance during development.