NVIDIA Warp Tutorial

This tutorial contains notebooks for learning NVIDIA Warp, an open-source Python framework for writing high-performance simulation and graphics code on CPUs and NVIDIA GPUs.

These notebooks can be run on NVIDIA Brev or Google Colab.

• Docker Images and Docker Compose files for creating Brev Launchables or running locally.

Brev Launchables of this tutorial should use:

• L40S, L4, or T4 instances.
• Crusoe or any other provider with Flexible Ports.

Notebooks

Notebook	Description	Colab
01. Introduction to NVIDIA Warp	Core Warp programming model, arrays, kernels, automatic differentiation, and a galaxy simulation capstone.
02. Ising Model	A 2D Ising model simulation built with Warp kernels and GPU arrays.
03. Navier-Stokes Solver	Builds a 2D Navier-Stokes solver using Warp kernels, tiled FFTs, and CUDA graph capture.
04. Differentiable Navier-Stokes Solver	Uses Warp automatic differentiation to optimize Navier-Stokes simulation inputs.

Layout

Notebook-specific images are stored under notebooks/images/ in subdirectories named for each notebook. This keeps assets organized as more Warp notebooks are added.

Additional Resources

• NVIDIA Warp documentation
• NVIDIA Warp GitHub repository
• Warp example gallery