triton-skills

What problem does it solve?

This Skill helps you write, debug, and optimize Triton and Gluon GPU kernels using local source code, tutorials, production kernel references, and compiler internals.

Core Features & Use Cases

• Kernel Development Guidance: Learn how to implement Triton and Gluon kernels for operations like matmul, fused attention, top-k, SwiGLU, and block-scaled computation.
• Performance Optimization: Find practical patterns for autotuning, persistent kernels, warp specialization, TMA, WGMMA, and tcgen05 to improve throughput and memory efficiency.
• Reference-Driven Troubleshooting: Use curated tutorials, quick references, and production implementations to debug API usage, layout issues, and numerical behavior.
• Use Case: If you need to build a high-performance FP8 attention kernel or tune a fused MoE matmul path, this Skill points you to the right source files and implementation patterns.

Quick Start

Ask for help implementing, debugging, or optimizing a Triton or Gluon kernel and include the target operation, tensor shapes, and hardware constraints.