croq-dsl-cuda

Generate CUDA kernel build scripts and profile performance with NVIDIA Nsight Compute.

Updated Apr 10, 2026

Search Tags:#ncu #performance profiling #cuda #nvcc #gpu optimization #kernel tuning

Ready to run

npx skills add https://github.com/LancerLab/croqtile-tuner --skill croq-dsl-cuda

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: croq-dsl-cuda
Source: https://github.com/LancerLab/croqtile-tuner/tree/main/.claude/skills/croq-dsl-cuda
Command: npx skills add https://github.com/LancerLab/croqtile-tuner --skill croq-dsl-cuda

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill simplifies the process of tuning CUDA kernels by providing standardized build and run scripts, environment validation, and performance profiling tools.

Core Features & Use Cases

Build Automation: Generate optimized CUDA kernel binaries with custom build scripts tailored for different shapes and models.
Performance Profiling: Utilize NVIDIA Nsight Compute (ncu) to collect detailed performance metrics on GPU kernels.
Use Case: A developer optimizing a deep learning operation can automate the compilation and profiling process to identify bottlenecks and improve throughput.

Quick Start

Use the croq-dsl-cuda skill to generate build and run scripts for your CUDA kernels and perform performance profiling with ncu.

Frequently Asked Questions about croq-dsl-cuda

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate CUDA kernel compilation and profiling?▼

Automating CUDA kernel compilation and profiling involves generating custom build scripts for different shapes and executing performance profiling with ncu to identify bottlenecks.

What is the best way to profile GPU kernels for deep learning operations?▼

The best way to profile GPU kernels is to use performance profiling tools like ncu to collect detailed metrics and identify bottlenecks in your deep learning operations.

Can I use custom build scripts for different CUDA kernel shapes and models?▼

You can generate optimized CUDA kernel binaries with custom build scripts tailored for different shapes and models to streamline your kernel tuning process.

Why do I need environment validation for GPU optimization?▼

You need environment validation for GPU optimization to ensure scripting automation and detailed performance analysis with NVIDIA tools function correctly on your high-performance GPU applications.

Does this approach support performance profiling with NVIDIA Nsight Compute?▼

This approach fully supports performance profiling with NVIDIA Nsight Compute (ncu) to collect detailed performance metrics on your GPU kernels for optimization.

When do I need automated build scripts for CUDA kernel tuning?▼

Automated build scripts for CUDA kernel tuning are needed when you want to streamline the process of building, executing, and profiling GPU code for high-performance applications.