skills.rest
NVIDIA
NVIDIAOFFICIAL
Explore all skills by this author →
WebsiteGitHub

adding-cutile-kernel

Automate adding cuTile kernel operators to TileGym with registration, backend, tests, and benchmarks.

796|82|Updated Nov 13, 2025
Search Tags:#testing#dispatch#gpu#kernel#cutile#tilegym
One-click install
npx skills add https://github.com/NVIDIA/TileGym --skill adding-cutile-kernel
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: adding-cutile-kernel
Source: https://github.com/NVIDIA/TileGym/tree/main/.claude/skills/adding-cutile-kernel
Command: npx skills add https://github.com/NVIDIA/TileGym --skill adding-cutile-kernel

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides an end-to-end workflow for adding a new cuTile GPU kernel operator to TileGym, including registration, backend implementation, exporting, tests, and benchmarking.

Core Features & Use Cases

  • End-to-end operator integration: from dispatch interface to cuTile backend, tests, and benchmarks.
  • Clear integration guidance: steps for wiring new op into ops.py, init.py, and test suites.
  • Reproducible workflow: ensures consistent scaffolding for future cuTile operator support.

Quick Start

Run the complete cuTile kernel addition workflow from steps 1 through 6 to add a new operator to TileGym.

Frequently Asked Questions about adding-cutile-kernel

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

FAQPage Schema
How do I add a new cuTile kernel operator to TileGym?▼

Adding a cuTile kernel operator to TileGym involves a complete end-to-end workflow: dispatch registration, backend implementation, exports, tests, and benchmarking. This ensures reproducible scaffolding for new operators.

What is the process for cuTile dispatch registration in TileGym?▼

Dispatch registration for cuTile in TileGym involves wiring the new operator interface into ops.py. This structured step ensures the operator is properly exposed for backend implementation and testing.

Does adding a cuTile GPU kernel require writing tests and benchmarks?▼

Yes, adding a cuTile GPU kernel requires writing tests and benchmarks. The TileGym integration workflow mandates including test suite steps and benchmarking to ensure reproducible and validated operator performance.

How do I wire a new cuTile op into TileGym exports?▼

Wiring a new cuTile op into TileGym exports requires updating __init__.py. This structured integration step ensures the operator is correctly dispatched, implemented in the backend, and accessible for testing.

What is the best way to standardize cuTile operator integration in TileGym?▼

The best way to standardize cuTile operator integration in TileGym is using an automated workflow covering dispatch, backend, exports, tests, and benchmarks. This ensures consistent, reproducible scaffolding for future operators.

Related Skills

Explore More Software Engineering Skills→
yo-stevenyo-steven

adding-cutile-kernel

Adds a cuTile GPU kernel operator to TileGm with full integration.

Community
Advanced
sayalinvidiasayalinvidia

tilegym-adding-cutile-kernel

Add a cuTile kernel operator to TileGym with dispatch registration and benchmarking.

Community
Advanced
NVIDIANVIDIA
796

monkey-patch-kernels-to-transformers

Monkey-patch TileGym CUDA kernels into HuggingFace Transformers models.

Official
Advanced
psengerpsenger
9

agentic-skeleton-dir-structure

Generate standardized project skeletons for agentic AI projects.

Community
Advanced