Query NVIDIA CUDA documentation for kernel development and debugging.
npx skills add https://github.com/m0at/claudemd --skill cuda-skill
Please help me install this Agent Skill. Skill: cuda-skill Source: https://github.com/m0at/claudemd/tree/main/skills/cuda-programming Command: npx skills add https://github.com/m0at/claudemd --skill cuda-skill
NVIDIA CUDA documentation is scattered across PTX ISA references, CUDA Runtime and Driver APIs, programming guides, and profiling guides. This Skill centralizes these resources (PtX, runtime/driver docs, Nsight Compute, Nsight Systems, best practices) to speed up lookup and improve debugging and optimization workflows.
Quickly locate CUDA PTX and runtime docs and Nsight references to accelerate kernel development.
High-intent search queries and answers about installing and using this skill.
Find CUDA Runtime API signatures by querying centralized documentation that covers PTX ISA, runtime, driver, and programming guides. This resource loads reference data on demand to streamline kernel development and API lookups.
Profile CUDA kernel performance by accessing Nsight Compute and Nsight Systems references for optimization tasks. The centralized documentation provides best practices for memory access and profiling workflows to improve debugging.
Search PTX ISA and CUDA Driver API references together through a centralized repository. This consolidated access speeds up lookup for kernel development by loading discovery data from frontmatter descriptions during activation.
CUDA documentation resources include best practices for memory optimization and profiling. References cover memory access recommendations from the programming guide to accelerate kernel debugging and performance tuning workflows.
Locate the API signature for cuModuleLoad by querying the CUDA Driver API documentation. The skill provides fast reference and search capabilities across runtime, driver, and PTX docs to streamline kernel development.
CUDA kernel debugging requires multiple sources because documentation is scattered across PTX ISA, runtime, driver, and profiling guides. Centralizing these resources speeds up lookup and improves optimization workflows.
GregsGreyCodeOptimize LLM inference on NVIDIA GPUs with TensorRT.
ar4mirezProvide CUDA kernel design and performance optimization best practices.
yo-stevenInstrument PyTorch training loops and CUDA kernels to measure latency and throughput.
dotruruLocate CUDA/PTX documentation and APIs in offline references.
