srush/GPU-Puzzles resource

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

12428 stars · 950 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

An interactive Jupyter notebook of programming puzzles that teaches beginner GPU/CUDA kernel programming using Numba's Python-to-CUDA mapping. Learners implement kernels directly in the notebook, designed to be run in Google Colab with a GPU runtime.

Use cases

• learn CUDA programming interactively
• practice writing GPU kernels in Python
• understand how deep learning GPU algorithms work
• teach parallel programming with puzzles
• get hands-on with Numba CUDA without writing C++

When to choose

• you know Python and want to learn GPU programming from scratch
• you want a hands-on, puzzle-driven introduction to CUDA kernels
• you want to build intuition for how GPU parallelism powers deep learning

When to avoid

• you need production CUDA code or performance tuning guidance
• you want a comprehensive CUDA reference or textbook
• you have no access to a GPU or Colab runtime

Facets

learning-resource · maturity stable

developer-tools gpu-computing machine-learning gpu-computing education machine-learning tutorials python browser cuda jupyter-notebook numba puzzles interactive-learning kernels gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="srush/GPU-Puzzles")