The LiHPC has 27 research engineers, one group leader assisted by three persons providing technical and administrative support. The LiHPC is currently made up of two teams, one in Computer Science for Simulation (section 27) and the other in Applied Mathematics (section 26).

The LiHPC is an applied research laboratory that covers all the activities necessary for numerical simulation :

• parallel programming languages for heterogeneous computers,
• Data flow and input/output management for codes,
• simulation code development platforms,
• the development of numerical methods and simulation codes,
• estimation of precision and numerical errors,
• exploitation of the large volumes of data from the codes (visualization and processing),
• geometry meshing and database management for codes,
• technology watch on new technologies (processors, accelerators, interconnect networks, software) and adaptation of simulation codes to those technologies.

Some of LiHPC’s work is integrated into software or codes for use in computing centers. LiHPC makes its software available according to open-source principles to promote and accelerate its research, in collaboration with national and international partners. It also promotes HPC computing at the national and international level.

The HPC scientific computing complex at the CEA facility in Bruyères-le-Châtel (DAM/Île de France) hosts one of Europe’s largest high-performance computing facilities, used for defence, industrial and research applications. This computing center is managed by the Department of Simulation Sciences and Information (DSSI).

It houses the large infrastructure operated by CEA DAM specifically for defence-related programmes, featuring the Atos/Bull EXA1 supercomputer (with a processing power of 36 petaflops) since December 2021. The centre’s unrestricted area also hosts two large infrastructures as part of the CEA’s TGCC (Very Large Computing Centre) facility. The first of these serves the needs of the European academic research community. At its core is the Atos/Bull Joliot Curie petascale supercomputer; the CPU time is provided to European research organisations under the Partnership for Advanced Computing in Europe PRACE programme, and also reserved with French research labs, via a civil society named GENCI. The second large infrastructure, designed to serve industry, is supported by the CEA’s Computing Centre for Research and Technology CCRT, which hosts since 2021 TOPAZE petascale supercomputer (8.8 petaflops), descendant of the Tera1000 class, used by industrial partners and CEA laboratories.

These operational computing centers are supported by a technological experimentation and expertise cluster, which provides the research and development expertise crucial to managing the complexity of large computing infrastructures.

The resulting computing complex is part of a broader effort to build a high-performance computing ecosystem. In keeping with this strategy, the CEA set up Ter@tec, an organisation that now has more than 80 industrial and academic partners.

High performance computing page on CEA site

The LiHPC is located at the Ter@tec Campus close to the TGCC, which facilitates interaction with other HPC actors.

The Laboratory has collaborative spaces to promote technical exchanges, offices, and access to the open computing resources of the DAM. LiHPC partners can be hosted in these premises for short/medium/long periods.