AmanBanik/Numba_Accelerated_Wolfworks_007 ? reverse-engineered prompt

Build me a polished learning module around GPU acceleration with Numba and CUDA, using Python notebooks. I want it to feel like a guided handbook that starts from the basics, then moves into grid and kernel ideas, memory handling, math examples like complex numbers and coordinate geometry, and finally physics simulations like N body, heat transfer, waves, and fluids.

Please make it easy to open and follow in Google Colab, with each lesson clearly separated and runnable on a GPU. Add simple explanations, code examples, and visual outputs where they help, so someone learning this for the first time can actually understand what’s going on. Also include the performance side, like streams, async work, pinned memory, and profiling, and keep the examples focused on real CUDA speedups.

If you need to look up current docs online for Numba or Colab details, go ahead and do that.