Goals: Students understand fundamental theoretical and practical concepts of advanced parallel algorithms and data structures, get to know established methods and algorithms, are familiar with issues of efficient implementations, are able to identify/formulate algorithmic problems in/for different application areas where parallel or distributed algorithms are used, are able to analyse new distributed and parallel algorithms as well as analysing their running time, and select appropriate algorithms for parallel and distributed applications -- students are able to apply parallel and distributed algorithms and data structures to real-world problems, and can objectively assess the quality of the results.

Content: Introduction to distributed and parallel algorithms, PRAM model, design and analysis of parallel and distributed algorithms, isoefficiency, UMA vs. NUMA, memory consistency for shared-memory, communication models (with and without network, fully interconnected with half duplex or full duplex, BSP), critical path lengths, parallel associative operations, reduction operations, matrix multiplication, broadcast operations, MPI basic toolbox, ranking, parallel sorting (multiway merge, quick sort, sample sort), prefix sums, all-to-all communication, map-reduce, list ranking, parallel graph algorithms (minimum spanning trees, connected components, shortest paths, graph partitioning), process mapping, communication-free parallel graph generation, parallel sampling algorithms.

Prerequisites: Recommended: Einführung in die Praktische Informatik (IPI), Programmierkurs (IPK), Algorithmen und Datenstrukturen (IAD), Lineare Algebra 1

Leistungspunkte: 8 LP

Duration: one semester

Teaching mode: 4 SWS lecture (in English), 2 SWS tutorial, homework assignments

Workload: 240h; thereof 90h lectures and tutorials, 15h exam preparations, 135h lecture wrap-up and homework

Passing the course: Successful participation in the exercises (at least 50\% of total achievable points) and passing an oral exam

Verwendbarkeit:
M.Sc. Angewandte Informatik,
M.Sc. Scientific Computing,
Lehramt