Intro to GPU programming

Harnessing the power of Graphics Processing Units (GPUs) can significantly accelerate computations, enabling solutions to problems previously considered intractable due to their computational complexity. This course provides a practical introduction to GPU programming, tailored for scientists, engineers, and programmers looking to leverage massively parallel computing resources effectively.

Key topics covered include:

• Introduction to GPU Architectures: Understand the fundamental design principles, strengths, and limitations of GPUs compared to traditional CPUs.

• CUDA Programming: Learn NVIDIA’s CUDA framework for writing high-performance GPU code, exploring kernels, memory management, and parallel computation.

• Accessing GPUs from Python: Discover how to leverage GPUs within the Python ecosystem using tools like Numba, CuPy, and PyCUDA.

• Hardware-Agnostic Programming: Explore cross-platform and hardware-agnostic frameworks like SYCL and WebGPU, enabling code portability and flexibility across different GPU architectures.

• Shader Programming: Gain insights into GPU shader programming, commonly used in graphics, visualization, and real-time processing applications.

Throughout the course, participants will engage in practical exercises and real-world examples to reinforce learning, ensuring a deep understanding of parallel programming concepts and best practices.

The course is adaptable and can be tailored to specific software stacks, programming environments, or application domains, accommodating participants from diverse scientific and engineering backgrounds.