Profiling tools for GPU;

Kernel performance metrics;

Profiling workflow demonstration;

Registers and memory hierarchy;

Global vs shared memory;

Memory access & optimisation strategies;

CUDA kernel structure;

Launch configuration (blocks/threads);

Kernel execution workflow;

Quiz #1;

JIT compilation concepts;

@cuda.njit for CPU;

Universal Functions (Vectorize)

Python performance optimisation basics;

Introduction to CUDA Architecture

Introduction to CuPy

Tensor operations and CPU <-> GPU array transfers.

NumPy vs CuPy for GPU acceleration;