OpenMP in the Scope of Real-Time Embedded Systems; Eduardo Quiñones talks in GdR SOC2 - IRT Saint Exupéry: "HPC for Embedded Systems"

Toulouse

The recent thematic day focused on high-performance computing in real-time embedded systems featured the participation of Eduardo Quiñones, Principal Investigator (PI) and coordinator of the European project Ampere from the Barcelona Supercomputing Center. His talk, titled "OpenMP in the Scope of Real-Time Embedded Systems," was scheduled from 13h30 to 14h05.

During his presentation, Quiñones discussed the performance requirements of advanced real-time embedded systems and the complexity of the parallel and heterogeneous processor architectures being used to deploy these systems. He highlighted the need for parallel programming models that provide performance, time predictability, and resilience, among other non-functional requirements. Quiñones focused on OpenMP, a parallel programming model traditionally used in high-performance computing, and its potential in adapting to real-time embedded systems.

The thematic day covered various aspects of high-performance computing and real-time embedded systems, starting with an introduction to the state-of-the-art architectures and hardware mechanisms for high-performance computing. Attendees were given an overview of industrial application needs, with an illustration in the automotive domain. The event addressed the capabilities and challenges posed by these hardware architectures for use in critical real-time embedded systems.

Further presentations delved into specific aspects, such as deterministic parallel computing models, security issues associated with high-performance architectures, and the implementation and exploitation of massive parallelism at the system and accelerator levels.

The event provided valuable insights into the evolving landscape of high-performance computing and its potential applications in real-time embedded systems, fostering discussions on the development and deployment of software on these platforms.

Check the event summary