PA177 High Performance Computing

Faculty of Informatics
Spring 2010
Extent and Intensity
4/0. 5 credit(s) (plus extra credits for completion). Type of Completion: zk (examination).
Teacher(s)
Prof. Thomas Sterling (lecturer), prof. RNDr. Luděk Matyska, CSc. (deputy)
doc. RNDr. Eva Hladká, Ph.D. (lecturer)
prof. RNDr. Luděk Matyska, CSc. (lecturer)
RNDr. Tomáš Rebok, Ph.D. (lecturer)
RNDr. Jan Fousek, Ph.D. (seminar tutor)
RNDr. Tomáš Golembiovský (seminar tutor)
Mgr. Roman Šustek (seminar tutor)
Guaranteed by
prof. RNDr. Václav Matyáš, M.Sc., Ph.D.
Department of Computer Systems and Communications – Faculty of Informatics
Contact Person: prof. RNDr. Luděk Matyska, CSc.
Prerequisites
The course is given in English only in a combination of real-time videoconferencing and assisted re-play of recorded lectures (in High Definition video quality). So very good fluency in English is necessary. The course includes practical exercises and home work (all in English), the knowledge of Linux computing environment is also required. A previous experience in using clusters and/or batch queue systems is helpful but not mandatory. Basic programming skills in C and Java are also beneficial.
Course Enrolment Limitations
The course is also offered to the students of the fields other than those the course is directly associated with.
The capacity limit for the course is 15 student(s).
Current registration and enrolment status: enrolled: 0/15, only registered: 0/15, only registered with preference (fields directly associated with the programme): 0/15
fields of study / plans the course is directly associated with
there are 18 fields of study the course is directly associated with, display
Course objectives
The goal of the course is to rapidly increase the number and capability of young scientists and engineers contributing to the application and development of high performance computing to address the needs of this nation and the welfare of its citizens through a highly practical treatment of all relevant aspects of the field. The course is interdisciplinary combining critical elements from hardware technology and architecture, system software and tools, and programming models and application algorithms with the cross-cutting theme of performance management and measurement. The course is hands-on with strong educational reinforcement through experimental exercises and quantifiable metrics. The course is multi-media for rapid and wie distribution employing state-of-the-art high-definition digital video and high bandwidth Internet networking technologies to support enhanced local classroom, geographically distributed real-time lectures, separately conducted classroom courses, and self-scheduled independent studies programs. To facilitate this high-impact college course a new textbook is being developed in conjunction with a web-site for distribution of course material and management of the student learning experience. NSF sponsorship will support the development of an Instructors Manual and the creation of the digital video lecture series library for pedagogical reinforcement through repetition and off-site independent studies.
Syllabus
  • Definition of the High Performance Computing, survey of current methods and a brief history of supercomputing.
  • Large scale applications, major computational problems requiring current and future generation supercomputers, resource requirements.
  • Enabling technologies, brief history of device technology, current used technologies.
  • Single node architecture and performance.
  • Parallel computer architecture and performance.
  • Commodity clusters as an example of an HPC system.
  • Benchmarking
  • Throughput computing and Condor.
  • MPI programming.
  • Performance monitoring, metrics, and measurements.
  • Parallel kernel algorithms and application design.
  • Domain specific programming environments.
  • Visualization.
  • System software.
  • Parallel I/O.
  • Beyond the Basics, advanced approaches like FPGA, streaming, and alternative programming models.
  • Towards the future (beyond Moore's Law, quantum computing, self healing systems).
Literature
  • ICS06 : 2006 National Conference on Supercomputing : June 28-July 1, Cairns, Queensland, Australia. New York: ACM Press, 2006, xii, 374. ISBN 1595932828. info
  • ICS05 : proceedings of the 19th ACM International Conference on Supercomputing : June 20-22, 2005 (Workshop Tutorials - June 19th), The Cambridge Marriott (Kendall Square) and The Stata Center (MIT), Cambridge, Massachusetts, USA. New York: ACM Press, 2005, x, 405. ISBN 1595931678. info
  • Optical supercomputing : first International Workshop, OSC 2008 : Vienna, Austria, August 26, 2008 : proceedings. Edited by Shlomi Dolev - Tobias Haist - Mihai Oltean. Berlin: Springer, 2008, 127 s. ISBN 9783540856726. info
  • More literature available on the web pages of the course
Teaching methods
Lectures given remotely via high definition streaming (including life interaction with the teacher), extensive home work that includes practical solution of given problems
Assessment methods
The course is given twice a week. It will include some home work with practical exercises, including setup and run of computer programs on a dedicated cluster. There will be several practical exercises during the semester, on average 2 more hours per week will be needed to complete all the assignments. The course is actually running at Louisiana State University (LSU) and the FI students will get the same load and treatment. Students are expected to attend all the lectures as only there a local lecturer will offer guidance and feedback (apart from the help with practical exercises). There will be a mid-term examination and a last written examination. According to the LSU university rules there may be just one possibility to take the exam, but this is still under negotiation. The home work and exercises will be part of the final evaluation. There is no required textbook for the course, all the relevant material will be provided to enrolled students at the end of lectures and via the course web pages (more detailed information and access codes will be given to enrolled students during the first lecture).
Language of instruction
English
Follow-Up Courses
Further comments (probably available only in Czech)
Study Materials
The course is taught annually.
The course is taught: every week.
Teacher's information
http://www.cct.lsu.edu/csc7600/
The course is also listed under the following terms Spring 2007, Spring 2008, Spring 2009.
  • Enrolment Statistics (recent)
  • Permalink: https://is.muni.cz/course/fi/spring2010/PA177