High Performance Computing M2

Objectives

This specialization is designed to provide scientific managers with a mastery of two major technological (r)evolutions:

• the increasingly critical use of numerical simulation, both in the industrial (design of new products, from ventilator blades to automobiles) and the research areas,
• the systematic use of (multicore) parallelism throughout the full range of processors.

One of the key characteristics of this specialization is that it provides graduates with a multidisciplinary background (applied mathematics and computing), which enables them to interact with experts in different fields. CEA DAM IdF is heavily involved in the project. The TER@TEC consortium, a European pole in High Performance Digital Simulation, also plays a considerable part in this specialized training program.

Program

The training objective is to develop advanced skills in the field of architecture and high performance parallel computers and software.

First Semester (30 ECTS)

• Core courses
  • Advanced compilation
  • High performance microprocessor architecture and code optimization
  • Software engineering for scientific computing
  • Performance evaluation
  • FGPA, GPU, architecture and programming
  • Project
  • Technical writing and speaking (English or French)
• Elective courses
  • Modelling methods in Biology, Modelling methods in Finance, Modelling methods in Fluid Mechanics,
  • Embedded systems architecture, High performance algorithms for signal processing.

Second Semester (30 ECTS)

• Internship