Home About Skills Experience Education Publications Contact GitHub LinkedIn

Benoît Gallet, Ph.D.

Postdoctoral Scholar

School of Informatics, Computing and Cyber Systems

Northern Arizona University

About


Postdoctoral scholar in Cybersecurity and Computer Science/Ecoinformatics at Northern Arizona University.

Cybersecurity: Porting Number Theoretic Transform (NTT) operations to GPU Tensor Cores for the Post-Quantum Cryptography (PQC) Kyber algorithm.

Ecoinformatics: Modelling tree water and bark beetle stresses using satellite imagery for the state of Arizona, in charge of the large-scale computing aspect.

Expertise in High Performance Computing (HPC), parallel and GPU computing, and algorithm optimizations.

Interested in data science and machine learning, familiar with TensorFlow and Keras.

Looking for job, willing to relocate.

F1 STEM OPT since June 2023, can be extended to Summer 2026.

Skills


Proficient with C, C++, Python, CUDA, OpenMP, OpenMPI

Familiar with Tensorflow and Keras

High-performance computing, parallel and GPU computing, algorithms, data structures, data analysis and clustering algorithms, algorithm optimizations.

Conducting research, presenting, public speaking.

French (native), English (bilingual), German (basic)

Experience


Postdoctoral Scholar
June 2023 - Present
Northern Arizona University, Flagstaff, USA
Cybersecurity
  • Porting and optimizing Number Theoretic Transform (NTT) operations of the Post-Quantum Cryptography (PQC) algorithm Kyber from the CPU to GPU Tensor Cores using CUDA.
Computer Science/Ecoinformatics
  • Part of a multi-disciplinary team, focusing on the scalability and performance of the project.
  • Designing and developing an algorithm to periodically retrieve satellite images for the state of Arizona and from multiple sources (PlanetLabs, USGS, NASA, and ESA), using Python.
  • Designing and developing an algorithm to periodically process new imagery, modeling and computing the trees water and bark beetle stresses of up to 40B pixels. Uses C++, and MPI for distributed computing and scalability.
PhD Student and Graduate Research Assistant
August 2018 - May 2023
Northern Arizona University, Flagstaff, USA
Dissertation: Efficient Euclidean Distance Calculations and Distance Similarity Searches: An Examination of Heterogeneous CPU, GPU and Tensor Core Architectures.
Worked on data analysis and clustering algorithms, with an emphasis on similarity searches and Euclidean distance calculations:
  • Developed GPU workload balancing optimizations (up to 9.7x speedup over prior GPU solution).
  • Developed a heterogeneous CPU-GPU algorithm (up to 5.5x speedup over prior GPU solution).
  • Developed a GPU algorithm using double precision Tensor Cores (up to 2.2x speedup over prior GPU solution).
CS Instructor
May 2022 - August 2022
Northern Arizona University, Flagstaff, USA
  • Instructor of record for the CS450 Introduction to Parallel Programming class.
  • Taught shared memory parallelism, including pthreads, OpenMP, and vectorization.
  • Average course evaluation: 3.86 / 4.
MSc Intern
April 2018 - September 2018
Université d’Orléans, Orléans, France and Northern Arizona University, Flagstaff, USA
GPU Kernel Performance Optimizations for Efficient Similarity Joins.
  • Proposed several optimizations for a GPU distance similarity searches algorithm using CUDA.
BSc Intern
April 2016 - June 2016
Université d’Orléans and National Center for Scientific Research, Orléans, France
GPU Detection of Pulse Radio Signals.
  • Ported C code detecting pulse radio signals from neutron stars with a radio telescope and using FFTs to the GPU using CUDA.

Education


PhD - Informatics and Computing
Northern Arizona University, Flagstaff, USA
August 2018 - May 2023
Efficient Euclidean Distance Calculations and Distance Similarity Searches: An Examination of Heterogeneous CPU, GPU and Tensor Core Architectures
Dissertation

MSc - Computer Science
Université d'Orléans, Orléans, France
September 2016 - September 2018
  • Minor in Nomadism, Intelligence and Security
  • With Honors

BSc - Computer Science
September 2013 - June 2016
Université d'Orléans, Orléans, France

Publications


Peer-Reviewed Publications (6)
  1. Optimization and Comparison of Coordinate- and Metric-Based Indexes on GPUs for Distance Similarity Searches
    Michael Gowanlock, Benoit Gallet, Brian Donnelly

    Proceedings of the International Conference on Computational Science (ICCS), Prague, Czech Republic, July 2023

    PDF DOI
  2. Leveraging GPU Tensor Cores for Double Precision Euclidean Distance Calculations
    Benoit Gallet, Michael Gowanlock

    Proceedings of the 29th IEEE International Conference on High Performance Computing, Data, and Analytics (HiPC), Bengaluru, India, December 2022

    PDF GitHub DOI
  3. Data-Intensive Computing Modules for Teaching Parallel and Distributed Computing
    Michael Gowanlock, Benoit Gallet

    11th NSF/TCPP Workshop on Parallel and Distributed Computing Education
    Proceedings of the IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

    PDF DOI
  4. Heterogeneous CPU-GPU Epsilon Grid Joins: Static and Dynamic Work Partitioning Strategies
    Benoit Gallet, Michael Gowanlock

    Data Science and Engineering, October 2020

    PDF DOI GitHub
  5. HEGJoin: Heterogeneous CPU-GPU Epsilon Grids for Accelerated Distance Similarity Join
    Benoit Gallet, Michael Gowanlock

    Proceedings of the 25th International Conference on Database Systems for Advanced Applications (DASFAA), Jeju, South Korea, September 2020

    PDF DOI GitHub
  6. Load Imbalance Mitigation Optimizations for GPU-Accelerated Similarity Joins
    Benoit Gallet, Michael Gowanlock

    IEEE High-Performance Big Data, Deep Learning, and Cloud Computing Workshop (HPBDC)
    Proceedings of the 2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), Rio de Janeiro, Brazil, May 2019 (Best Paper Award)

    PDF DOI GitHub

Posters (2)
  1. Optimizing GPU-Accelerated Similarity Joins Addressing Data-Dependent Workloads
    Benoit Gallet, Michael Gowanlock

    IEEE International Parallel and Distributed Processing Symposium (IPDPS), PhD Forum, Rio de Janeiro, Brazil, 2019

    PDF
  2. Exploring The Design-Space of GPU-Efficient Similarity Self-Join Kernels
    Benoit Gallet, Michael Gowanlock

    High-Level Parallel Programming and Applications (HLPP), 11th International Symposium, Orléans, France, 2018

    PDF

Awards and Other
  • Best Paper Award
    Load Imbalance Mitigation Optimizations for GPU-Accelerated Similarity Joins
    IEEE High-Performance Big Data, Deep Learning, and Cloud Computing Workshop (during IPDPS'19)
  • IEEE TCPP Travel Grant, $1,100
  • Selected for the IEEE IPDPS'19 PhD Forum

Contact


School of Informatics, Computing and Cyber Systems
1295 S Knoles Dr
Flagstaff, AZ, United States
Cubicle #301E-4