events

What Happens Inside a GPU

A talk hosted by Houston Machine Learning that explains what a GPU does while a kernel runs, built around an analogy of a busy train station: cargo handlers stand in for threads, terminals for streaming multiprocessors, and cargo for data m

Starts

2026-11-06T20:00:00+00:00

Organizer

Houston Machine Learning

Location

Online event

A talk hosted by Houston Machine Learning that explains what a GPU does while a kernel runs, built around an analogy of a busy train station: cargo handlers stand in for threads, terminals for streaming multiprocessors, and cargo for data moving through the memory hierarchy. The session covers how GPUs schedule work, feed Tensor Cores, and overlap data transfers with computation, and asks why larger tiles are not always faster and why workloads and GPU generations call for different execution strategies. No prior kernel-writing experience is required.