Complete searchable markdown conversion of NVIDIA's PTX ISA 9.1, CUDA Runtime API 13.1, and CUDA Driver API 13.1 documentation with Claude Code skill for GPU development.
git clone https://github.com/technillogue/ptx-isa-markdown.gitThis skill converts NVIDIA's PTX ISA 9.1 and CUDA API documentation from fragmented HTML pages into 640+ organized markdown files, enabling fast offline lookup via grep or Claude search. It includes complete instruction set reference, Runtime and Driver API function documentation, data structure definitions, and profiling/debugging workflows. The Claude Code skill provides PTX instruction lookup, CUDA API reference, TensorCore operation details, and profiling patterns using nsys and ncu. Developers working with CUDA can ask Claude questions about register layouts, error codes, context management, or memory operations and instantly retrieve relevant documentation with section references.
Copy the cuda_skill directory to ~/.claude/skills/cuda. The skill activates automatically for CUDA work. Ask Claude questions like 'What's the register fragment layout for WGMMA m64n16k16?' or 'How do I disable TMA swizzling?' Claude searches the local documentation and provides answers with references.
Look up PTX instructions and register fragment layouts for WGMMA operations
Find CUDA Runtime API functions, parameters, and error code meanings
Query CUDA Driver API for low-level context and memory management
Reference TensorCore (WMMA, WGMMA, TMA) operation details
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/technillogue/ptx-isa-markdownCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Convert the following PTX ISA 9.1 documentation to searchable markdown format. Include all technical specifications, instructions, and examples. Ensure the output is well-structured with clear headings and bullet points. Focus on the following sections: [SECTIONS]. Also, integrate Claude Code skill for CUDA development where applicable. Use the provided data: [DATA].
# PTX ISA 9.1 Documentation ## Overview PTX ISA 9.1 is a virtual machine instruction set architecture for NVIDIA GPUs. It enables developers to write parallel programs that can be executed on NVIDIA's CUDA architecture. ## Key Features - **Parallel Execution**: Supports parallel execution of threads. - **Memory Hierarchy**: Includes shared, global, and constant memory. - **Atomic Operations**: Supports atomic operations for synchronization. ## Example Code ```cuda // Example of PTX ISA 9.1 code .mad.lo.s32 %r1, %r2, %r3, %r4; // Multiplies and adds two 32-bit integers ``` ## Claude Code Integration For CUDA development, Claude Code can assist in: - **Code Generation**: Automatically generate PTX code from high-level CUDA code. - **Optimization**: Suggest optimizations for PTX code. - **Debugging**: Identify and fix issues in PTX code. ## Conclusion PTX ISA 9.1 is a powerful tool for developers working with NVIDIA GPUs. Its integration with Claude Code enhances the development process, making it more efficient and effective.
AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
Automate security compliance and monitor real-time security posture seamlessly.
Automate your spreadsheet tasks with AI power
Agentic AI Workflow platform
Connected workspace for docs, wikis, and projects
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan