The PARKAS project focuses on the design, semantics and compilation of high-level languages which allow to go, all the way down, from a parallel deterministic specification to a target embedded code that may execute on parallel (multi-core) architectures. The team bases its research on the theory and practice of synchronous languages, typed functional languages and modern compilation techniques (polyhedral compilation) to obtain provably safe and efficient code. The team develops languages and compilers (e.g., Zelus, ReactiveML, contributions to GCC and LLVM) used as vehicles for validating, communicating and transferring research results. The language Lucid Synchrone, for example, has been the basis of SCADE 6 developed by Esterel-Technologies and used for implementing critical embedded software.
Design, semantics and implementation of synchronous languages.
Extensions of synchronous languages to target new applications (e.g., N-synchrony and relaxed synchrony for computer-intensive video applications, mixed continuous/discrete signals for modeling physical systems, dynamic creation for large scale simulations).
Type systems, static analysis of synchronous programs.
Code generation with proved efficiency and safety for sequential and parallel (multi-core) processors.
ANSYS/Esterel-Technologies; Dassault-Systèmes; ARM; Kalray
We are always on the lookout for excellent, highly motivated students and young researchers for research internships, PhD theses, postdoc, research engineer, and researcher positions.