Skills
.Work. You
Rest

converting-cutile-to-triton

Official

Convert cuTile kernels to Triton with automation.

Software Engineering#gpu#triton#autotune#tma#cutile#kernel-conversion#ct.load
AuthorNVIDIA
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Converts cuTile kernels to Triton, enabling automated porting, debugging, and optimization of GPU kernels between cuTile and Triton.

Core Features & Use Cases

  • Automated cuTile → Triton translations with guidance on 2D+ loads, TMA usage, and layout flags
  • Dual-layout handling with separate kernels and META-based grids to preserve performance across transpose-like patterns
  • End-to-end workflow guidance from pre-flight analysis to correctness verification and benchmarking

Quick Start

Follow the step-by-step workflow to convert a cuTile kernel to Triton and validate correctness and performance.

Dependency Matrix

Required Modules

None required

Components

references

💻 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: converting-cutile-to-triton
Download link: https://github.com/NVIDIA/TileGym/archive/main.zip#converting-cutile-to-triton

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.