Skip to main content

IDRIS courses

IDRIS offers courses in High Performance Computing (HPC) and Artificial Intelligence (AI). These courses are free of charge for CNRS and university staff, who can register via the website https://cours.idris.fr/. They are also open to employees of companies, EPIC and EPSTs, who can register through CNRS Formation Entreprise.

If you would like further information, feel free to contact us:
📥 e-cours@idris.fr

Check the schedule of upcoming sessions:
🗓️ Upcoming sessions

Course catalog

HPC​

MPI (Message Passing Interface). The objective of this course is to enable participants to develop parallel programs by implementing the MPI message passing library. IDRIS, Orsay (91). Duration: 4 days. Prerequisites: Fortran or C, Unix basics.

OpenMP (Open Multi-Processing). This course is intended for anyone wishing to parallelize an application for a shared-memory multicore architecture. IDRIS, Orsay (91). Duration: 3 days. Prerequisites: Fortran or C.

OpenMP & MPI: Intensive week. This course is intended for users who want to develop parallel programs using the MPI library and get an introduction to parallelization with OpenMP. IDRIS, Orsay (91). Duration: 5 days. Prerequisites: Unix basics and knowledge of Fortran or C

Introduction to hybrid MPI/OpenMP programming. This course is intended for anyone wishing to get started with hybrid MPI/OpenMP programming, particularly for optimal use of supercomputers. IDRIS, Orsay (91). Duration: 2 days. Prerequisites: Fortran 90/95 or C, solid knowledge of MPI and OpenMP is required

Introduction to GPU programming with OpenACC/OpenMP target. This course is intended for anyone wishing to understand hybrid accelerated architectures (GPU) and use OpenACC and OpenMP 5.0 directives to parallelize code. IDRIS, Orsay (91). Duration: 3 days. Prerequisites: Fortran 90/95 or C, basic knowledge of OpenMP and/or MPI (optional).

SIMD Vectorisation (Simple Instruction Multiple Data). This course is intended for users who want to get started with SIMD vectorization (Single Instruction Multiple Data) and apply these concepts to concrete examples. IDRIS, Orsay (91). Duration: 1 day. Prerequisites: Unix basics and Fortran.

HPC Debugging. This course is intended for anyone wishing to learn the DDT debugging tool in the context of parallel programming used in HPC and to be able to efficiently debug scientific compute codes. IDRIS, Orsay (91). Duration: 1 day. Prerequisites: OpenMP, MPI.

Introduction to the PETSc library. This course is intended for anyone wishing to learn the basic skills needed to get started with the PETSc library and to build and solve simple examples of partial differential equations. IDRIS, Orsay (91). Duration: 2 days. Prerequisites: Fortran or C, Linux basics and text editor usage, knowledge of linear algebra and MPI.

Introduction to the ScaLAPACK and MAGMA libraries. This course is intended for anyone wishing to parallelize linear algebra codes on single-GPU and multi-GPU systems (MAGMA) as well as on distributed-memory MIMD parallel computers (ScaLAPACK). IDRIS, Orsay (91). Duration: 2 days. Prerequisites: Fortran or C, basic knowledge of linear algebra and MPI programming (optional)

Modern Scientific C++. This course is intended for anyone wishing to update their knowledge of the language’s latest features, with the goal of writing scientific computing and data processing software applications. IDRIS, Orsay (91). Duration: 3 days. Prerequisites: Knowledge of legacy C++ syntax (before C++11).
Introduction to Kokkos programming. This course introduces the basics of Kokkos on CPU and GPU. It is intended for beginners who want to develop a first program using fundamental data structures, parallel loops, and parallel reductions. IDRIS, Orsay (91). Duration: 3 days. Prerequisites: Basic knowledge of C++, CMake, and Linux

Artificial Intelligence​

Practical Introduction to Deep Learning (IPDL). This course aims to provide participants with fundamental knowledge of deep learning, with an emphasis on understanding core concepts and essential best practices. IDRIS, Orsay (91). Duration : 2 days. Prerequisites : basic knowledge of algebra, statistics, and Python.

Deep Learning Architectures (ArchDL). The objective of this course is to familiarize participants with a variety of neural network architectures suited to different types of data such as images, audio, text, or graphs. IDRIS, Orsay (91). Duration: 2 days. Prerequisites: Python syntax, basics of algebra, statistics, and deep learning.

Deep Learning optimized for Jean Zay (DLO-JZ). This course aims to review the main techniques for optimizing a deep learning training, with a focus on scaling up to a supercomputer. IDRIS, Orsay (91). Duration: 4 days. Prerequisites: Python, deep learning, and basic knowledge of PyTorch.
LLM specialization (SpeLLM). This course is aimed at anyone wishing to adapt an LLM via fine-tuning, set up a specialization and optimization environment, and explore multimodality and agent-based systems. IDRIS, Orsay (91). Duration: 4 days. Prerequisites: Python, Deep Learning, PyTorch

Support sessions​

Workshop Jean Zay. This session is intended for all users, both beginners who want to quickly get started working on Jean Zay and more experienced users aiming to optimize their applications. IDRIS, Orsay (91). Duration: 1–2 days. Prerequisites: an existing account or a pending account request on Jean Zay

Deep Learning Kickoff Workshop (ALDL). The objective of this workshop is to help kick-start a deep learning project by providing a project review with an AI engineer from IDRIS. IDRIS, Orsay (91). Duration : Half-day session. Prerequisites: completion of an introductory deep learning course