Develop GPU kernels using Triton syntax with Python-like DSL.
npx skills add https://github.com/mindspore-ai/akg --skill triton-syntax
Please help me install this Agent Skill. Skill: triton-syntax Source: https://github.com/mindspore-ai/akg/tree/main/akg_agents/examples/run_skill/skills/triton-syntax Command: npx skills add https://github.com/mindspore-ai/akg --skill triton-syntax
Triton语言与编程模式为GPU核函数的开发提供简化入口，使GPU编程变得更像Python，降低学习成本并提高开发效率。
Write and run a simple Triton kernel that adds two vectors.
High-intent search queries and answers about installing and using this skill.
You can write a GPU kernel using Python by leveraging the Triton DSL, which provides Python-like syntax to streamline GPU programming and reduce implementation friction for custom operations.
The block programming model in Triton supports efficient parallel computing and cache management, allowing developers to build GPU-accelerated workloads like vector adds and matrix multiplications with automatic optimization.
Yes, you can use Triton to develop custom GPU operators for matrix operations, as it is specifically designed to build GPU-accelerated workloads including matrix multiplications and vector calculations.
The best way to start writing a simple Triton kernel is to implement a vector addition, which demonstrates the Python-based Triton DSL syntax and automatic optimization features for GPU programming.
No, you do not need to manage cache manually, because Triton features automatic optimization and a block programming model that handles efficient parallel computing and cache management for your GPU kernels.
flashinfer-aiExtract GPU kernel schemas from SGLang models into deduplicated Definition JSON files.
plurigridGenerate deterministic colors across CPU and GPU with SplitMix64 PRNG.
scottgl9Record CUDA kernel calls and tensor states to diagnose GPU errors.
jstzwjWrite CUDA kernels for NVIDIA GPU computing with memory management.
