High Performance Computing

  • Help
  • Find
  • Facebook
  • Twitter
Study

Choose UPPA

    • A multi-campus university

      The Université de Pau et des Pays de l’Adour (UPPA) is a network of five campuses.

      Read more

    • International student life

      UPPA offers its students international mobility through multiple partnerships.

      Read more

Keywords

High Performance Computing

Objectives

Discover how and why numerical methods performances are varying with respect to their implementations. We focus on large scale computing architecture and parallelism.

 

Recommended prerequisite

M1 MMS mandatory courses and L3 numerical analysis

Number of hours

  • CM : 27
  • TD : 12

Form of assessment

First session

Second session

Continuous assessment: 30%

Final examination: 70%

Final examination length: 3 hours

Remedial Examination: 100 %

Remedial Examination length: 3 hours

 

Syllabus

1- Usual programming languages

2- Usual numerical methods for EDP models

3- Sequential High Performance Computing : algorithm complexity, diagnosis, conservation

4- Different memory architectures

5- Domain decomposition methods

6- Parallel computing 1 : MPI standard, message passing parallelism

7- Parallel computing 2 : OpenMP standard, shared memory parallelism

8- Hybrid computing : CPU and GPU, arrays mapping

9- An example : analogic neural network

In brief

ECTS credits 4

Number of hours 39

Level of study Master degree level

Contact(s)

Organizational unit

Administrative contact(s)

Secrétariat de Mathématiques - Brigitte GAUBERT

Email : brigitte.gaubert @ univ-pau.fr

Places

  • Pau