Julien Bigot

Julien Bigot

Auto-generated profile, based on publicly available information

Julien Bigot

Université Paris-Saclay, UVSQ, CNRS, CEA, Maison de la Simulation, 91191, Gif-sur-Yvette, France

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2021
Verified
Conference Paper
Closed access
AGAmal Gueroudji
Amal Gueroudji
JBJulien Bigot
Julien Bigot
BRBruno Raffin
Bruno Raffin
A widening performance gap is separating CPU performance and IO bandwidth on large scale systems. In some fields such as weather forecast and nuclear fusion, numerical models generate such amounts of data that classical post hoc processing is not feasible anymore due to the limits in both storage capacity and IO performance. In situ approaches are attractive to bypass disk accesses in these cases and fully leverage the HPC platform. They are however often complex to set up and can require to re-develop parallel versions of the analysis from scratch. In this paper we propose a hybrid model that is well suited for in situ workflows that combine regular simulations and irregular analytics. Our model couples the bulk synchronous parallel paradigm for simulation with a distributed task-based one for analysis. This reduces complexity and leverages the best of each of these two powerful paradigms. We validate the model with a prototype, called DEISA, that supports coupling MPI parallel codes with analyses written using Dask. This implementation requires minimal modifications of both the simulation and analysis codes compared to their post hoc counterpart. It give access to an already existing rich ecosystem to be used in situ such as the parallel versions of Numpy, Pandas and scikit-learn. Experiments in configurations up to 1024 cores show that DEISA can improve the simulation wallclock time (excluding analysis) by a factor up to 3 and the total experiment (including analysis) hour. core cost by a factor of up to 5 compared to parallel post hoc with plain Dask while requiring the modification of only two lines of python code, three of YAML, and none at all in a C simulation code already instrumented with PDI Data Interface.

9 September 2025 in Journal of Open Source Software
Verified
Journal Article
Open access
PDF file
EBEmily Bourne
Emily Bourne
VGVirginie Grandgirard
Virginie Grandgirard
YAYuuichi Asahi
Yuuichi Asahi
JBJulien Bigot
Julien Bigot
PDPeter Donnel
Peter Donnel
PVPauline Vidal
Pauline Vidal
Gyselalib++ provides the mathematical building blocks to construct kinetic or gyrokinetic plasma simulation codes in C++, simulating a distribution function discretised in phase space on a fixed grid. It relies on the Discrete Domain Computation (DDC) library (Padioleau et al., 2025) to statically type the discretisation dimensions, thus preventing many common sources of errors. Via DDC, Gyselalib++ also leverages the Kokkos framework (Trott et al., 2022), ensuring performance portability across various CPU and GPU architectures. The library provides a variety of tools including semi-Lagrangian advection operators, quadrature rules, and solvers for elliptic and hyperbolic partial differential equations (PDEs). The majority of the operators are designed to work on non-orthonormal coordinate systems; those that don't use the static typing to raise compiler errors preventing their misuse.

1 May 2017
Verified
Conference Paper
Open access
OAOlivier Aumage
Olivier Aumage
JBJulien Bigot
Julien Bigot
HCHelene Coullon
Helene Coullon
CBChristian Pérez Bernal
Christian Pérez Bernal
JRJerome Richard
Jerome Richard
This paper studies the feasibility of efficiently combining both a software component model and a task-based model. Task based models are known to enable efficient executions on recent HPC computing nodes while component models ease the separation of concerns of application and thus improve their modularity and adaptability. This paper describes a prototype version of the COMET programming model combining concepts of task-based and component models, and a preliminary version of the COMET runtime built on top of StarPU and L2C. Evaluations of the approach have been conducted on a real-world use-case analysis of a subpart of the production application GYSELA. Results show that the approach is feasible and that it enables easy composition of independent software codes without introducing overheads. Performance results are equivalent to those obtained with a plain OpenMP based implementation.

12 November 2023
Verified
Conference Paper
Open access
AGAmal Gueroudji
Amal Gueroudji
JBJulien Bigot
Julien Bigot
BRBruno Raffin
Bruno Raffin
RRRobert Ross
Robert Ross
In situ workflows are inescapable to fully leverage exascale architectures. They can be complex to build, however, because simulation and data analytics come from two different software ecosystems with their own paradigms and programming models. This work extends the deisa bridging model between MPI+X simulations and distributed task-based analytics; it introduces the concept of external tasks to support the description of analytics graphs spanning multiple timesteps ahead of time while improving scalability. This new approach leads to a straightforward support for contracts between the simulation and analytics graph to limit the data transferred to that actually analyzed in a given execution. We implement this approach using Dask and MPI and evaluate it using an end-to-end in-transit workflow that uses an unsupervised ML model for dimensionality reduction. We compare our work with plain Dask postprocessing and with the previous version of deisa. Our work performs better, up to × 7 for the simulation and × 3 for the analytics compared with deisa, and is × 18 less costly compared with plain Dask—all of these with similar development efforts.

15 November 2025
Verified
Conference Paper
Open access
EMEtienne Malaboeuf
Etienne Malaboeuf
KOKévin Obrejan
Kévin Obrejan
MPMathieu Peybernes
Mathieu Peybernes
JBJulien Bigot
Julien Bigot
EBEmily Bourne
Emily Bourne
VGVirginie Grandgirard
Virginie Grandgirard

30 July 2025 in Journal of Open Source Software
Verified
Journal Article
Open access
PDF file
YAYuuichi Asahi
Yuuichi Asahi
TPThomas Padioleau
Thomas Padioleau
PZPaul Zehner
Paul Zehner
JBJulien Bigot
Julien Bigot
DLDamien Lebrun-Grandie
Damien Lebrun-Grandie
Kokkos-fft provides a unified, performance-portable interface for Fast Fourier Transforms (FFTs) within the Kokkos ecosystem (C. Trott et al., 2021). It seamlessly integrates with leading local FFT libraries including FFTW, cuFFT, rocFFT, and oneMKL. Designed for simplicity and efficiency, kokkos-fft offers a user experience akin to numpy.fft for in-place and out-of-place transforms, while leveraging the raw speed of vendor-optimized libraries. A demonstration solving 2D Hasegawa-Wakatani turbulence with the Fourier spectral method illustrates how kokkos-fft can deliver significant speedups over Python-based alternatives without drastically increasing code complexity, empowering researchers to perform high-performance FFTs simply and effectively.