Architectures and Middleware for Scientific Computing

The lecture discusses architectures and middleware for scientific computing and big data applications. It consists of two parts: in the first part we focus on GPU programming and in the second part we investigate advanced topics from the field of cluster computing and parallel programming.


Topics

  • Part 1: GPU Programming
    • GPU hardware
    • programming models for GPUs: CUDA, OpenCL, OpenACC

  • Part 2: Cluster Computing
    • selected MPI functions: MPI one-sided, processor topologies
    • parallel file systems
    • Architectures for Big Data: From Hadoop to Kafka, Key-Value Stores

Requirements

Basic C programming skills are necessary. The bachelor lecture "Konzepte paralleler Programmierung" is recommended.


Lecturer

Prof. Dr. Bettina Schnor


Modules

  • Master Computational Science: 7010
    • 552511 - Vorlesung
    • 552521 - Übung
    • 552501 - Prüfung

  • Master Data Science
    • INF-DSAM4A: Advanced Infrastructures and Software Engineering A
    • INF-DSAM4B: Advanced Infrastructures and Software Engineering B

Room/Dates

The lecture is given in room 03.04.0.02 as a block course at the following dates:


  • Monday, 14.10.2019, 16:15 - 17:45 h
  • Monday, 21.10.2019, 10:00 - 18:00 h (all the day), room 03.04.1.02
  • Tuesday, 22.10.2019, 09:00 - 17:00 h (all the day), room 03.04.1.02
  • Monday, 04.11.2019, 16:15 - 17:45 h
  • Monday, 11.11.2019, 16:15 - 17:45 h
  • Monday, 18.11.2019, 16:15 - 17:45 h
  • Monday, 25.11.2019, 16:15 - 17:45 h
  • Monday, 02.12.2019, 16:15 - 17:45 h
  • Monday, 09.12.2019, 16:15 - 17:45 h
  • Monday, 13.01.2020, 16:15 - 17:15 h Exam test, room 03.04.0.02 (as usual)
  • Monday, 20.01.2020, 16:15 - 17:45 h Final project presentation, room 03.04.0.02 (as usual)
  • Monday, 27.01.2020, 16:15 - 17:45 h Final project presentation
News
Notenspiegel 13.01.2020 Notenspiegel
First lecture 14.10.2019, 16:15 - 17:45 h Universität Potsdam, room 03.04.0.02