PhD student in Computer Science. My research focused on HPC with the topic : “Energy efficiency, Exascale and Complex Systems”. I work at the Univerrsity of Reims Champagne-Ardenne and the CReSTIC Laboratory EA3804, France.
I work in collaboration with the ROMEO HPC Center.
What will be the structure of the tomorow supercomputer to reach exascale ? GPUs, Xeon Phi ? InfiniBand and NVLink ?
How using these accelerator and optimize the programmation according to this massively parallel architecture ?
Computational problems, Langford used as a showcase. Irregular problems as benchmarks like Graph500 using BFS on huge graphs.
I worked on astrophysic simulation at the Los Alamos National Laboratory during Summer 2016. For two months and a half during a Co-Design Summer Shool. This project brings my interest on physics particles problems and their acceleration using GPU.
BFS Traversal on Multi-GPU Cluster Many-Core Approaches to Combinatorial Problems: case of theLangford problem Parcours de grands graphes sur architecture hybride CPU/GPU Towards the Parallel Resolution of the Langford Problem on a Cluster of GPU Devices Domain Partitioning and Problem Space Representations for Compact Binary Mergers Massively Parallel Resolution of Combinatorial Problems on MultiGPU ClustersBegining of my PhD thesis which will last three years in the CReSTIC laboratory of the University of Reims Champagne-Ardenne, Reims, France
My Master's degree specialized in "Development of parallel and distributed applications".
With honors and first in graduating class.
My Bachelor's Degree in computer science is a general formation concerning computer science theory, development, network and hardware considerations.
With honors
As the continuation of my work during my first intership we worked on another method to beak a new instance of the academic problem of Langford.
Using all the ROMEO supercomputer we were able to compute L(2,27).
Then I proposed a bibliography on massively parallel resolution of SAT problems.
This intership was my first contact with HPC. I worked on the ROMEO supercomputer using using MPI + OpenMP + CUDA runtime.
The subject was the Langford problem and a general scheme to solve combinatorial problems represented as CSP.
Cette dernière semaine a été bien remplie ! Un article a rédiger et beaucoup de travail au laboratoire. Mais un soir en rentrant Ann m'a proposé d'aller faire un petit…
Enfin ! Après un super Week-end de promenades, direction le laboratoire Mardi 31 pour le premier jour. Je me fais déposer par Blake au bâtiment "Otowi" ! Direction la "New Employes…