Ad­van­ced Com­pu­ter Ar­chi­tec­ture (MSc CS, CE)

Advance Information

This is an overview of the contents of the lecture "Advanced Computer Architecture" that will be offered in Winter Term 2023/24.

Course Information

  • Title: Advanced Computer Architecture
  • Lecturer: Prof. Dr. Christian Plessl
  • Lecture Number: L.079.05724
  • Teaching Language: English

Format

  • 2h lecture
  • 1h theoretical exercise
  • 3h practical exercises/lab per week

Times

  • Lecture: Thusdays, 14:00 - 16:00
  • Exercise: Thursdays, 16:00 - 17:00
  • Lab: Group 1: Monday, 11:00 - 14:00; Group 2: Tuesday, 11:00 - 14:00; Group 3: Wednesday, 8:00 - 11:00 

Lab 3 is likely to be removed, so if possible please join lab group 1 or 2!

Content Overview

The course Advanced Computer Architecture teaches concepts and methods used in modern processor architecture to exploit the available parallelism at the levels of instructions, data and threads.

After attending the course, students are able …

  • to explain principles of modern memory hierarchies
  • to analyze different levels of parallelism
  • to assess the suitability of different architectural concepts and thus
  • to evaluate developments in computer architecture

Who Should Take this Course?

This course is mandatory for students of the Master program Computer Engineering, and an elective for students of the Master program Computer Science.

Although there are no formal prerequisites for taking this course, we expect that students have completed a Bachelor-level computer architecture course. Such a course should provide basic knowledge in instruction set architectures, assembly language programming, processor performance evaluation, data path and controller design, pipelining, and memory hierarchies. 

We strongly recommend this self-assessment test to find out whether you should take this course.  

Tentative List of Topics

The course covers the following topics:

  • Fundamentals of computer architectures
  • Memory hierarchy design
  • Instruction-level parallelism
  • Data-level parallelism: Vector, SIMD and GPU architectures
  • Thread-level parallelism
  • Warehouse-scale computers
  • Domain-specific architectures

Materials

  • Lecture slides
  • Panda Course

Further Reading