Oli Tissot

  • Since the end of October 2019, I am a software developer at MathWorks working on the Parallel Computing Toolbox in the Parallel Algorithms team.
    From February 2019 to October 2019, I was a post-doctoral research associate at Johannes Gutenberg Universität (Mainz, Germany), working in the group of Markus Bachmayr.
    Before that, I was a Ph.D. student at the Laboratoire Jacques-Louis Lions (Sorbonne university) as part of the Inria team Alpines under the supervision of Laura Grigori. My Ph.D. has been funded through the H2020 project NLAFET.

    • Research interests

    • Numerical Linear Algebra
    • High-performance Computing
    • Tensor computations
    • Sparse Matrix Computations

    Software

    Bocop toy test: a mouse escapes a maze.

    preAlps (2016 - 2019)

    Developper of the parallel version(s) of the E(nlarged) C(onjugate) G(radient) method in C and MPI inside the preAlps library.
    Features
    • Matrix and preconditioner free (Reverse Communication Interface).
    • Very light: it only needs BLAS and LAPACK (Intel MKL).
    • Scalability assessed up to 16,384 cores on elasticity matrices.
    • Documentation and examples.
    Github project

    Bocop (2014 - 2015)

    Developper et maintainer of Bocop, a C++ optimal control toolbox.
    • Windows/Mac/Linux compatible.
    • User friendly GUI (based on Qt).
    • HJB approach that allows to solve stochastic problems.

    Academic projects (2011 - 2014)

    M.Sc.
    • 2D Helmholtz and wave equations solver (C++).
    • 2D Navier-Stokes equations solver (C++).
    • Parallel low-rank decomposition (C and MATLAB).
    Github project
    B.Sc.
    • Building and programing a ball collector robot.

    Research

    Ph.D. thesis

    • OT, Iterative methods for solving linear systems on massively parallel architectures, pdf.

    Publication

    • Laura Grigori, OT, Scalable linear solvers based on Enlarged Krylov subspaces with dynamic reduction of search directions, SIAM Journal on Scientific Computing, 41(5), C522–C547, 2019, pdf.

    Technical reports

    • Laura Grigori, OT, Enlarged Conjugate Gradient with adaptive reduction of search directions, pdf (old version).
    • Simplice Donfack, Laura Grigori, OT, NLAFET Deliverable 4.5: Integration, pdf.
    • Maksims Abalenkovs, Simplice Donfack, Jack Dongarra, Iain Duff, Laura Grigori, Stojce Nakov, Jan Papez, OT, Mawussi Zounon, NLAFET Deliverable 5.2: Software integration, pdf.
    • Simplice Donfack, Laura Grigori, OT, NLAFET Deliverable 4.4: Performance evaluation, pdf.
    • Simplice Donfack, Laura Grigori, OT, NLAFET Deliverable 4.3: Prototype software, pdf.
    • Simplice Donfack, Laura Grigori, OT, NLAFET Deliverable 4.2: Analysis and algorithm design, pdf.
    • Frédéric Bonnans, Daphné Giorgi, Benjamin Heymann, Pierre Martinon, OT, BocopHJB 1.0.1 – User Guide, pdf.

    Talks

    • Enlarged Conjugate Gradients for reducing communication, Seminar of the Scalable Algorithms group at Sandia National Laboratories, Albuquerque, USA (June 2018).
    • Enlarged Conjugate Gradient method for reducing communication, SIAM PP18, MS 55, Tokyo, Japan (March 2018).
    • Enlarged GMRES for reducing communication, SIAM PP18, MS 28, Tokyo, Japan (March 2018). I filled in for Hussam Al Daas.
    • Enlarged Krylov subspace methods for reducing communication, PASC 17, MS 45, Lugano, Switzerland (June 2017).
    • Enlarged GMRES, SIAM CSE17, MS254, Atlanta, USA (March 2017). I filled in for Hussam Al Daas.
    • Adaptive enlarged Krylov conjugate gradient, DD24, MS01, Longyearbyen, Norway (February 2017).
    • Iterative methods for solving linear systems on supercomputers, Junior Seminar, Inria Paris (December 2016).
Address
Matrix House
Cambridge Business Park
10 Cowley Park
CB4 0HH Cambridge, UK
Contact
otissot (at) mathworks.com