Develop, debug, and optimize Triton and Gluon GPU kernels with tutorials and examples.
npx skills add https://github.com/dotruru/claudemd --skill triton-skill-dotruru
Please help me install this Agent Skill. Skill: triton-skill Source: https://github.com/dotruru/claudemd/tree/main/skills/triton Command: npx skills add https://github.com/dotruru/claudemd --skill triton-skill-dotruru
This Skill helps developers write, debug, and optimize Triton and Gluon GPU kernels by providing practical tutorials, source references, and hands-on examples in one guided package.
Start with 01-vector-add.py to initialize a basic Triton kernel.
High-intent search queries and answers about installing and using this skill.
Writing and debugging Triton GPU kernels starts with basic examples like vector addition. You can use guided tutorials to learn kernel development step by step, progressing to more complex patterns like matmul and reduction operations.
Converting CUDA kernels to Triton involves translating CUDA-style logic into Triton's programming model. This Skill provides structured references and practical examples to help you translate and optimize your existing GPU kernel code.
Autotuning in Triton optimizes GPU kernel performance by searching through configuration parameters. You can use provided tutorials to learn how to apply autotuning and profiling techniques for performance analysis in both research and production environments.
Yes, Triton supports FP8 and FP4 quantization in GPU kernel development. The Skill covers exploring these quantization methods alongside Tensor Memory Architectures to help you optimize and scale your kernel implementations.
Yes, it provides a curated set of kernel references and implementation patterns specifically for matmul, reduction, and attention. These examples guide you through common GPU kernel development scenarios and optimization strategies.
Yes, you need access to the Triton repository content via the provided paths to use this Skill effectively. The tutorials and source references rely on accessing the repository to demonstrate kernel development, debugging, and profiling.
m0atDevelop, debug, and autotune Triton and Gluon GPU kernels for ML workloads.
guqiong96Develop and optimize Triton and CUDA JIT kernels for diffusion models.
involvexDebug frontend code in the browser with breakpoints and console inspection.
spacecat2002Add and validate lightweight CUDA JIT kernels in SGLang's jit_kernel module.
