Universität Wien

052112 VU Numerical High Performance Algorithms (2023W)

6.00 ECTS (4.00 SWS), SPL 5 - Informatik und Wirtschaftsinformatik
Continuous assessment of course work
Moodle

Registration/Deregistration

Note: The time of your registration within the registration period has no effect on the allocation of places (no first come, first served).

Details

max. 25 participants
Language: English

Lecturers

Classes (iCal) - next class is marked with N

  • Tuesday 10.10. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 12.10. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 17.10. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 19.10. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 24.10. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Tuesday 31.10. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Tuesday 07.11. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 09.11. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 14.11. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 16.11. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 21.11. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 23.11. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 28.11. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 30.11. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 05.12. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 07.12. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 12.12. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 14.12. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 09.01. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 11.01. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 16.01. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 18.01. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 23.01. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG
  • Thursday 25.01. 13:15 - 14:45 Hörsaal 2, Währinger Straße 29 2.OG
  • Tuesday 30.01. 13:15 - 14:45 Seminarraum 6, Währinger Straße 29 1.OG

Information

Aims, contents and method of the course

Know and understand selected advanced numerical high performance algorithms (for example, divide-and-conquer eigensolver, GMRES, least squares solver, QR algorithm, communication-avoiding linear solver, etc.) for large and very large problems in computational science as well as in computational data science and machine learning. Understand the interdependencies between algorithm, problem data, implementation of the algorithm, hardware, performance and accuracy. Understand basic techniques for analysis, implementation and optimization of numerical high performance algorithms. Implement and evaluate your own implementations of selected algorithms. To some extent it may be possible to accomodate for special interests of the students in class.

PLEASE NOTE: It is currently planned to hold the course on-site (in presence).

Assessment and permitted materials

One or two homework exercises (with theoretical and practical components - implementation, experimentation, analysis), presentation and discussion of an assigned paper from the literature, and an individual semester project (involving literature research, implementation, experimentation, analysis), whose results will be presented in class and documented in written form (short project report, presentation slides).

Minimum requirements and assessment criteria

The maximum possible score is 100 points: 25 for the homework exercise(s), 25 for the paper presentation, 50 for the semester project (25 for the presentation, 25 for the report). At least 50 points are required for passing the course. Moreover, at least half of the available points for the semester project have to be achieved for passing the course.

Attendance of classes is in general strongly recommended. Attendance is required for the following classes: discussion of homework exercise(s), paper presentations, project presentations.

Examination topics

There is no separate exam. Grading takes into account the components mentioned above (homeworks, paper presentation, semester project) as well as the participation in class.

Reading list

Slides presented in class, literature references given on the slides.

Demmel: Applied Numerical Linear Algebra
Golub & van Loan: Matrix Computations
Trefethen and Bau, III: Numerical Linear Algebra

Association in the course directory

Module: HPA APS

Last modified: Mo 09.10.2023 21:07