Parallele Programmierung

Die Veranstaltung wird in Deutsch mit englischsprachigen Unterlagen durchgeführt.

The rise of multi-core and many-core systems and clusters requires that software engineers develop software which is executed concurrently and parallely. The course covers common problems, best-practices, general architectural concepts, frameworks and languages which target the development of concurrent and parallel programs.

Specifically, we will cover:

• Definitions and foundational concepts
• Parallelism in multi-core single systems:
• Threads (Java)
• Futures (Java)
• ThreadPools / Executor (Java)
• Coroutines
• Safe Concurrent Programming: Ownership, Moves, Borrowing, Type Safety, Channels (Rust)
• Actor-Modell (Akka)
• MapReduce

Lecture and Exercises:

The course will contain significant practical aspects. Small programming exercises will be part of the lecture accompanied by weekly assignments.

Learning objectives:

The learning objectives of the course are the following:

Knowledge and understanding:

• explain the common problems arising from concurrency and parallelism and possible solutions
• explain different frameworks to exploit parallelism and how they can be used
• explain how language paradigms support the development of concurrent applications
• explain the following concepts: threads, processes, interprocess-communication, multi-agent systems,

Skills and abilities:

• avoid common mistakes by applying best practices
• use frameworks to develop parallel applications on multi-core systems and clusters
• use different languages and frameworks by applying knowledge about general concepts

Judgement and approach:

• select appropriate concepts, languages and frameworks for developing concurrent and parallel applications