GPU Kernel Baseline — Development agent for Claude Code
name: gpu-kernel-baseline description: GPU kernel baseline implementation expert.
How to install GPU Kernel Baseline
Installs to ~/.claude/agents/alibaba-atrex-kernel-agent-gpu-kernel-baseline.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/alibaba/atrex-kernel-agent/HEAD/agents/gpu-kernel-baseline.md -o ~/.claude/agents/alibaba-atrex-kernel-agent-gpu-kernel-baseline.md Restart Claude Code, or start a new session, for it to be picked up.
What GPU Kernel Baseline does
name: gpu-kernel-baseline description: | GPU kernel baseline implementation expert. Learns the target framework from gpu-wiki, implements a correct baseline GPU kernel (V0), validates correctness, records performance, and produces all Stage 1 deliverables. Use when the user provides PyTorch logic and asks to build a baseline kernel for profile-driven optimization. tools: Read, Grep, Glob, WebSearch, WebFetch, Write, Bash
Role Definition
You are a GPU kernel baseline implementat
Alternatives in Development
- Kernel Cuda Specialist 14.5k ★
- Sample Renderer — 각 디자인 팩으로 실제 PPT 슬라이드/웹 페이지 샘플을 생성하고, 비교용 baseline 출력도 함께 생성해 preview.png를 산출하는 렌더링 엔지니어 188 ★
- Ascend Kernel Developer — Ascend kernel 开发专家 Agent,通过 TileLang 设计表达和 AscendC 落地完成算子优化任务 39 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Fk Kernel Engineer
Writes and integrates GPU kernels (Triton, TileLang, CuTe DSL, CUDA C++, CUDA graphs, torch.compile) under can
Metal Expert
Read-only Metal reviewer. Use when reviewing custom GPU rendering, MTKView, shaders, compute kernels, command
Kernel Cute Specialist
Kernel Tileir Specialist
Strategy Critic
PROACTIVELY use this subagent immediately after the user writes or revises any strategy artifact: Strategy Blo
Sap Basis Consultant
SAP Basis 장애 상황을 체계적으로 라우팅하고 진단하는 한국어 전문 에이전트. ABAP 덤프(ST22), Work Process 행(SM50/SM66), Transport 실패(STMS), R
Related Skills
Triton GPU Kernel Optimization
Write high-performance tiled Triton GPU kernels with autotune, grouped tile ordering, stride-based addressing,
Peridot
Peridot v1.6.0 [STABLE] Sovereign AI Kernel. Offline, GPU accelerated Qwen framework: a 27B model fully on an
Claudes C Compiler
Claude Opus 4.6 wrote a dependency-free C compiler in Rust, with backends targeting x86 (64- and 32-bit), ARM,