Part 3: GPU Computing: CuPy and PyTorch Tensors

Chapter 11: CuPy — NumPy on the GPU

Intermediate~150 min

Learning Objectives

  • Port NumPy code to CuPy with minimal changes for GPU acceleration
  • Perform GPU-accelerated linear algebra and FFT operations with CuPy
  • Implement custom CUDA kernels in CuPy using ElementwiseKernel and RawKernel
  • Implement efficient Kronecker matrix-vector products on GPU without forming the full matrix

Sections

Prerequisites

💬 Discussion

Loading discussions...