Parallel and High Performance Computing
(ehem. "Parallel Computing: Grundlagen und Anwendungen")
Vorlesung mit Übung im Wintersemester 2018/19
Prof. Dr. D. Kranzlmüller,
Dr. K. Fürlinger
This course will be held in English!
Welcome to the course webpage for Parallel and High Performance
for winter-term 2018/19 at LMU Munich. Here you will be
able to find details on the lecture and the accompanying practical lab
- There will be no tutorial on 17.01.19. The next tutorial will be held on 24.01.19.
- On Friday, Jan. 11 2019 at 11:00am, Andrey Semin from Intel will give a guest lecture on Application driven architecture design of SuperMUC-NG.
- The written exam will take place on Tuesday, Feb. 12, 2019,
at 12:30 pm, in Room M 218, Geschwister-Scholl-Platz 1. Please
register via UniWorX!
Parallel computing is concerned with using multiple compute units
to solve a problem faster or with higher accuracy. Historically, the
main application area for parallel machines is found in engineering
and scientific computing, where high performance computing (HPC)
systems today employ tens- or even hundreds of thousand compute
The application area for parallel computing has, however,
expanded recently to essentially include all areas of information
technology. Virtually all servers, desktop, and notebook systems, and
even smartphones and tables are today equipped with CPUs that contain
multiple compute cores. In each case, the potential for these systems
can only be fully realized by explicit parallel programming. As such
understanding the benefits, challenges, and limits of parallel
computing is increasingly becoming a "must have" qualification for IT
This course addresses the increasing importance of parallel and high
performance computing and is covering three interwoven areas:
Parallel hardware architectures, parallel algorithm
design, and parallel programming. The successful student
will be able to identify potentials for parallel computing in various
application areas, judge the suitability of contemporary hardware
architectures for a parallel computing problem and understand
efficient implementation strategies using modern parallel programming
The lecture is partially based on material that has been developed at
UC Berkeley and which has been funded by the US National Science
Foundation. The course slides will be made available for download by
the date of the lecture and will be in English.
The course is intended for both bachelor and master students of
computer science and related fields. More formally, in German: Die
Vorlesung richtet sich an Studenten der Informatik
bzw. Medieninformatik (Diplom) nach dem Vordiplom sowie an Studenten
der Informatik, Bioinformatik bzw. Medieninformatik (Bachelor, Master)
im Rahmen der vertiefenden Themen der Informatik. Für Vorlesung und
Übung werden 6 ECTS-Punkte vergeben.
- Lecture: Friday 9:30 - 12:00, Oettingenstr. 67, Room B001. First
lecture: Friday, October 19, 2018, starting at 9:15 in Room B001
- Lab exercise: Thursday 14:00 - 16:00 (ct), Geschw.-Scholl-Pl. 1 (F),
Room F 007. First lab: October 25, 2018, starting at 14:15 in Room F 007
- Please keep an eye on News for any short-term
changes and announcements.
- Friday January 11, 2019, 11:00am:
Guest lecture by Andrey Semin, Intel
Title: Application driven architecture design of SuperMUC-NG
Abstract: The presentation will provide details about
SuperMUC-NG - a new supercomputer deployed at Leibniz-Rechenzentrum in
Garching: the fastest supercomputer in Germany and among 91 from in
the EU countries (according to Top500 listing in November 2018) and
occupying 8th position in Top500 worldwide. However, the Linpack
benchmark was not the main workload that determined features of
SuperMUC-NG! SuperMUC-NG was designed to deliver the best performance
on wide range of practically useful scientific applications: there
were over 40 different workloads that drove the architecture decisions
and selection of main components. In the presentation we will describe
in detail some of the tools and methods used to optimize system
architecture to deliver high application performance. We will cover
the approaches to performance estimations, modelling and
extrapolations, as well as discuss design choices and the compromises
considered during design work.
Bio: Andrey Semin is a Principal Engineer in Intel
Architecture, Graphics and Software Group at Intel Corporation,
specializing in power and performance optimization and modelling of
high-performance computing (HPC) applications. He supports the leading
European high-performance computing users, helping them to deploy new
and innovative HPC solutions to grand-challenge problems. Andrey's
background includes extensive experience working with leading HPC
software and hardware vendors. Andrey has been instrumental in
developing HPC industry innovations improving energy efficiency from
data center to applications. Andrey graduated from Moscow State
University in 2000, specializing in possibility theory and its
applications for physical experiment analysis. He is the author of
over a dozen papers and patents in the area of application tuning and
energy efficiency analysis, and co-authored a book "Optimizing HPC
Applications with Intel Cluster Tools". He was honored with an Intel
Achievement Award (IAA) in 2018: highest award given to an individual
- The written exam (duration: 90 min.) for the lecture will
take place on Tuesday, Feb. 12, 2019, at 12:30 pm, in Room
M 218, Geschwister-Scholl-Platz 1.
The lecture is accompanied by a lab exercises to deepen the
understanding of topics covered in the lecture. High performance
computing systems hosted at the Leibniz Supercomputing Center will be
made available to the students. Worksheets for the lab exercises will
be made available on UniWorX.
Lecture slides will be made available chapter-by-chapter through this
Paul E. McKenney (Ed.): Is Parallel Programming Hard, And, If So, What Can You Do About It?(online) (link)
Ananth Grama et al.:Introduction to Parallel Computing (2nd Ed.) (link)
David Culler and Jaswinder Pal Singh: Parallel Computer Architecture, A Hardware / Software Approach (link)
John Hennessy and David Patterson: Computer Architecture a Quantitative Approach (5th Ed.) (link)
Georg Hager and Gerhard Wellein: Introduction to High Performance Computing for Scientists and Engineers (link)
Barbara Chapman et al.: Using OpenMP (link)
Ruud van der Pas, Eric Stotzer and Christian Terboven Using OpenMP - The next Step (link)
William Gropp, Ewing Lusk, Anthony Skjellum: Using MPI (link)
William Gropp, Torsten Hoefler, Ewing Lusk: Using Advanced MPI (link)
Please register for the course on UniWorX
and/or after the
lecture and lab exercises.
The repeat exam will for the course will be held as an oral exam on
one of the following listed dates. Please select a free slot and send
an email to Karl.Fuerlinger@nm.ifi.lmu.de to schedule an exam.
All exams will take place in room EU 102, Oettingenstr. 67
- Wed Apr. 10, 10:00: C. G.
- Wed Apr. 10, 10:30: K. D.
- Fri Apr. 12, 13:00: R. K.
- Fri Apr. 12, 13:30: L. W.
- Fri Apr. 12, 14:00: D. B.
- Tue Apr. 16, 9:00: Ch. N.
- Tue Apr. 16, 10:00: S. S.
- Thu Apr. 18, 13:00: Th. J.
- Thu Apr. 18, 13:30: A. J.
- Thu Apr. 18, 14:00: Th. K.
- Thu Apr. 18, 14:30: A. A.
- Thu Apr. 18, 15:00: M. H.