Research Scientist, Inria
Permanent member of the KerData research team at INRIA Rennes – Bretagne Atlantique Research Center and IRISA.
Inria Rennes Bretagne – Atlantique
Campus Universitaire de Beaulieu, 35042, Rennes
Office: D172 (orange level)
Phone: +33 (0)2 99 XX XX XX
My main current research interests are related to I/O and storage on large scale distributed infrastructures: clouds, HPC systems, converged infrastructures for HPC and Big Data analytics.
- SuperCompCloud at SC’22: 6th Workshop on Interoperability of Supercomputing and Cloud Technologies (Nov. 2022)
- Workshop co-chair
- PASC’22: Minisymposium “Storage Systems at Extreme Scale for Data-Centric Workflows” (June 2022)
François Tessier has been a Research Scientist at Inria in Rennes, within the KerData team since November 2020. He has been focusing there on I/O and storage challenges in a context of HPC/Cloud convergence. Before that, he worked two years as a Computational Scientist at ETH Zürich within CSCS (Swiss National Supercomputing Center) located in Lugano, Switzerland where he worked on dynamic provisioning of storage resources for HPC applications and large-scale workflows. Previously, he spent two years and a half as a postdoctoral appointee at Argonne National Laboratory, IL, USA where he addressed the problem of data movement optimization for I/O and storage. He received a Ph.D in Computer Science in 2015 from University of Bordeaux, France, under the supervision of Emmanuel Jeannot and Guillaume Mercier. His Ph.D thesis focused on topology and affinity-aware application placement and load balancing.
A list of publications can be found on Google Scholar. Below is a short list of the most representative papers:
- J. Monniot, F. Tessier, M. Robert, G. Antoniu, “StorAlloc: A Simulator for Job Scheduling on Heterogeneous Storage Resources” in The 20th International Workshop on Algorithms, Models and Tools for Parallel Computing on Heterogeneous Platforms (HeteroPar 2022), Glasgow, UK, 2022
- F. Tessier, M. Martinasso, M. Chesi, M. Klein, and M. Gila, “Dynamic provisioning of storage resource: a case study with burst buffers” in 2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), New Orleans, LA, USA, 2020
- F. Tessier, P. Gressier, and V. Vishwanath, “Optimizing data aggregation by leveraging the deep memory hierarchy on large-scale systems” in Proceedings of the 2018 International Conference on Supercomputing (ICS’18), Beijing, China, 2018
- F. Tessier, V. Vishwanath, and E. Jeannot, “TAPIOCA: an I/O library for optimized topology-aware data aggregation on large-scale supercomputers” in 2017 IEEE International Conference on Cluster Computing (CLUSTER), Honululu, HI, USA, 2017
- E. Jeannot, E. Meneses, G. Mercier, F. Tessier, and G. Zheng, “Communication and Topology-aware Load Balancing in Charm++ with TreeMatch” in 2013 IEEE International Conference on Cluster Computing (CLUSTER), Indianapolis, IN, USA, 2013
- E. Jeannot, G. Mercier, and F. Tessier, “Process placement in multicore clusters: algorithmic issues and practical techniques” in IEEE Transactions on Parallel and Distributed Systems, vol. 25, no. 4, pp. 993–1002, 2014
- StorAlloc: StorAlloc is a simulator of a storage-aware job scheduler. Through our simulator, we want to demonstrate how storage resources can be allocated the same way as computing resources on large-scale systems equipped with heterogeneous storage spaces.
- TAPIOCA: Topology-aware data aggregation C++ library for I/O intensive applications. The repository also contains a set of C++ and Fortran benchmark. Transfer actions of TAPIOCA into HDF5 are in progress.
- TreeMatch: Project initiated by Emmanuel Jeannot and Guillaume Mercier. TreeMatch is an algorithm and a library computing an affinity-aware placement of processing entities. The goal of TreeMatch is to minimize the communication cost between processes and reduce the overall time to solution of parallel and distributed applications.
- Dynamic Storage Resource Provisioning: Tool designed to provide, on-demand and through the job scheduler, a set of isolated intermediate storage resources along with a containerized storage system deployed on the fly on top of it. The current implementation supports three data managers: a parallel file system, a high-performance object store, and a highly scalable database.
Supervision of research activities
- Since Nov. 2021 – Julien Monniot. Co-supervision with Gabriel Antoniu. Topic: Dynamic provisioning of intermediate storage resources across hybrid HPC/Cloud infrastructures
- 2021 – Matthieu Robert (Université de Rennes): Co-supervision with Gabriel Antoniu. Topic: Algorithms for dynamic provisioning of storage resources on supercomputers