Registration:: https://www.eventbrite.ca/e/1015486787317

Learn to accelerate your Python functions on CPU and GPU using Numba.

Python isn't optimized for high-performance computing and can't be used “as is” on graphics processors. You can overcome these shortcomings using software libraries such as CUDA Python (PyCUDA) and Numba, which combine the ease of Python development with the power of a compiled language that targets both CPUs and GPUs.

In this workshop, you will gain insights into using these tools to create simple GPU programs and to understand the core principles of GPU programming.

Plan:

• Why write for GPUs?
• Understanding GPU architecture.
• What is CUDA?
• Python + CUDA: Your first PyCUDA code.
• GPUArray: A very useful PyCUDA library.
• Numba: Just-in-time compilation.
• Numba: Your first code.
• Numba functions and compilation modes.
• Numba + CUDA.
• Practical Session: Learning to program the GPU with PyCUDA and Numba.