Embedded Machine Learning

Course Overview

On the application side, HPC includes, among others, the computation, modeling, and simulation of complex systems from the areas of astrophysics, particle physics, biology, genetics, quantum chemistry, computational fluids, and weather and climate research. Such workloads have huge requirements with regard to computing power, which (still) cannot be satisfied completely. In this course, we will look at the question of how computer architectures can support such challenging applications in the best way possible. Key aspects are highly parallel hardware architectures like clusters, messaging-based communication, and software environments. Furthermore, we will shortly review scalable training of machine learning models in the second part of the course. The exercises will mainly contain practical work, including the development of own applications.

Contents

• Challenges of HPC
• Message passing
• Parallel programming
• Optimizations
• Workload characteristics
• Interconnection networks
• ML training

Requirements

Recommended is solid knowledge of C/C++ and the basics of computer architecture.

Notes

• Course starts Oct 17 9:00 c.t.
• Start of exercise is to be negotiated with the teaching assistant.
• Room is OMZ/INF350 basement, U014. Enter the building from the east. If you don’t see a ZITI sign, you are probably at the wrong place.