Jean-Philippe HALIMI

Jean-Philippe HALIMI

Research Engineer, University of Versailles

En poste chez University of Versailles

Précédents : University of Delaware, Exascale Computing Research Center


Précédents : University Of Versailles, Université Versailles Saint Quentin, Lycée Jean Rostand


    En résumé

    Master 2 MIHP student (High Performance Computing) at Ecole Centrale Paris and University of Versailles, specialized in optimizing techniques for programs ; compilation optimizations, vectorization, parallelism... Programming : - C/C++ (OpenMP/MPI), x86 assembly, CUDA - Web development (Javascript/HTML5/CSS3/PHP) I am interested in optimizations techniques using static and dynamic performance evaluation. I also have been involved in energy consumption related work.



Research Engineer

Chez University of Versailles

De octobre 2012 à aujourd'hui
Worked on a DVFS controller called FoREST which aims to reduce energy consumption at both processor and machine levels. FoREST is a tool that evaluates the impact of frequency transition on energy, using an innovative energy prediction model that is based on a short offline profiling.


Chez University of Delaware

De avril 2012 à octobre 2012
Developed a runtime based on a fine-grain program execution model which consists in a hybrid approach between Von Neuman and Dataflow models. Studied runtime characterization, bottlenecks identification using common benchmark programs such as Matrix Vector product, BFS or a Stencil program.

University Of Versailles, Versailles

Informatique a Haute Performance (HPC), University of Versailles

De 2010 à 2012
Computer Engineering - Parallel Programming, Computer Architecture, Program Characterization, GPU Architectures/Programming, Compilers, Graph Theory, Matrix Algorithms, Linear Programming


Chez Exascale Computing Research Center

Prise en compte et mesure de la problématique énergétique des processeurs Intel utilisés dans les futurs supercalculateurs des centres de calcul intensifs. Réflexion et implémentation de solutions permettant d'y parvenir.


  • Assembly
  • Compilation
  • CUDA
  • Energy
  • GPU
  • High Performance Computing
  • Openmp
  • Voir toutes les compétences (12)

Centres d'intérêt

  • Footing
  • Going out with friends
  • Video Games