Parallel Program Engineering (IN2310)

Prof. Dr. Michael Gerndt, Prof. Dr. Martin Schulz

 

Dates:Block course, 30.04.-04.05.2018, Frauenchiemsee
First meeting:30.04.2018
ECTS:5
Language:English
Type:Lecture, 4VÜ
Moodle course:Click here
Registration:By Email: gerndt@in.tum.de, First-come-first-served

 

 

PPE Group Fotos

Parallel Program Engineering 2018
Parallel Program Engineering 2016

Organization

The lecture will be given as a block lecture at Kloster Frauenwörth at Chiemsee by Prof. Gerndt and Prof. Schulz. We will stay at Frauenwörth for 5 days. The exercises will be projects with parallel programming tools and will be mixed with the lectures.

Background on parallel programming and/or parallel architectures is strongly recommended, although we will give an introduction in MPI, OpenMP, batch schedulers etc. It is also required that you bring your own laptop.

Register per Email (gerndt@in.tum.de) no later than September 30th. At most 12 students can participate. Students are accepted on a first-come-first-served basis.

For the stay at Frauenwörth there will be the following costs for you:

  • Cost for meals 180 € to be paid to the monastery

The oral exam will be held after the lecture. The dates can be arranged.

Schedule

Monday, 30.04.2018 9:04-9:58 Regional train Meridian to Salzburg from
München Ostbahnhof track 8
Exit at Prien am Chiemsee.
10:15-10:25 Chiemseebahn to Prien harbor
10:30-11:00 Boat to Fraueninsel
12:00 Lunch at Klosterwirt
13:30-17:30 Lectures
18:00 Dinner at Klosterwirt
Tuesday - Thursday 8:00 Breakfast
9:00 - 12:00 Lectures
12:00 Lunch
13:30 - 17:30 Lectures
18:00 Dinner
Friday 8:00 Breakfast
9:00 - 12:00 Lectures
12:00 Lunch
13:00 - 13:30 Boat to Prien
13:40 Chiemseebahn
14:09 - 15:05 Train to Munich

Contents

The course will present tools and tool infrastructures for parallel programming. The students will be able to apply the tools, to design new tool concepts, and to evaluated different implementation methods.

Structure

  • Introduction to parallel programming, programming models and languages, applications, and the parallelization approach.
  • Software development process focusing on aspects and requirements for parallel and high performance applications. Development environments supporting the orchestration of parallel programs.
  • Concepts and tools for engineering parallel programs focusing on:
    • Domain decomposition
    • Debugging
    • Performance analysis
    • Performance modeling and prediction
    • Application tuning
    • Performance engineering workflows
    • Tool development infrastructures

Practical Lab Course

In summer semester, a master lab course is given on Cloud Computing. Students work in groups and implement a scalable cloud application.