Post-Doctoral Research Visit F/M Task-based Cardiac Electrophysiology Simulation on multi-GPU Heterogeneous Clusters

Télécharger l'offre au format PDF
Le descriptif de l’offre ci-dessous est en Anglais

Type de contrat : CDD

Niveau de diplôme exigé : Thèse ou équivalent

Fonction : Post-Doctorant

Niveau d'expérience souhaité : De 3 à 5 ans

A propos du centre ou de la direction fonctionnelle

The Inria center at the University of Bordeaux is one of the nine Inria centers in France and has about twenty research teams.. The Inria centre is a major and recognized player in the field of digital sciences. It is at the heart of a rich R&D and innovation ecosystem: highly innovative SMEs, large industrial groups, competitiveness clusters, research and higher education players, laboratories of excellence, technological research institute...

Contexte et atouts du poste

International Relation Department Postdoctoral Position

Every year Inria International Relations Department has a few postdoctoral positions in order to support Inria international collaborations. The postdoctoral contract, within the context of the agreement signed between Inria and Simula Research Laboratory in Oslo, Norway, will have a duration of 24 months. The default start date is November 1st, 2025 and not later than January, 1st 2026. The postdoctoral fellow will be recruited by the Inria Research Center at the University of Bordeaux in France, but it is recommended that the time is shared between France and Norway  (please note that the postdoctoral fellow has to start his/her contract being in France and that the visits have to respect Inria rules for missions)

Task-based Cell-level Cardiac Electrophysiology Simulation

Simulating cardiac electrophysiology at the resolution of individual cells is a key challenge, as coarser-grained and cell-averaged simulations fail to capture the effects of pathological or aging cardiac tissues with sufficient realism. Cell-resolved simulation at such a fine-grained level necessitates the efficient exploitation of the computing power of contemporaneous HPC platforms equipped with accelerators. However, doing so comes with computer science challenges for HPC application developers such as deciding when and how to map application kernels onto these heterogeneous computing resources, handling data transfers, replication and consistency management between discrete accelerator memory spaces, between distributed nodes or ensuring the overlapped progression of data transfers in concurrence with computations. To address these issues, task-based parallel programming models, together with their associated runtime systems, such as the StarPU runtime system developed by Inria Team STORM, have shown the benefits of their ability at performance portability on HPC platforms.

The HPC Department of Simula has recently developed a primitive version of a new cell-resolved simulator of cardiac electrophysiology, named DEMI. This new simulator advances Simula's previous parallel cell-based simulation software, which was based on regular computational meshes and the finite difference numerical method. The new DEMI simulator adopts the finite element numerical method and can thus model realistic geometries of irregular shapes. To efficiently handle huge-scale computations that are required by cell-based simulations of
cardiac electrophysiology, there is an urgent need to port the DEMI simulator to cluster platforms of multiple GPU nodes. A standard CPU-to-GPU code porting will pose a major programming challenge. Thus, the task-based parallelization and runtime system StarPU is expected to provide a considerably more programmer-friendly approach.
 
Ongoing Collaboration
Simula’s HPC Department and Inria Team STORM are jointly involved in the European project MICROCARD-2 CoE (2025 - 2027, 30-month) in which they work on the OpenCARP cardiac electrophysiology simulation code, together with other partners of the project. In this respect, relevant scientific findings and experiences from this post-doc would be transferable to the MICROCARD-2 Project. The Post-doc researcher will be well posed to interact and collaborate with other consortium members of the MICROCARD-2 CoE, who are working on related research topics.
This current collaboration follows the joint participation of Simula and STORM to the MICROCARD Project (2020-2024) and to the associate team MAELSTROM. In consequence, Simula's HPC department will provide support in supervising and collaborating with the Post-doctoral researcher.
 
 

Mission confiée

Recruitement Guidelines

Candidates for International Relation Department postdoctoral positions are recruited after the end of their Ph.D. or after a first post-doctoral period: for the candidates who obtained their PhD in the Northern hemisphere, the date of the Ph.D. defense shall be later than September 1, 2022; in the Southern hemisphere, later than April 1, 2022. In order to encourage mobility, the postdoctoral position must take place in a scientific environment that is truly different from the one of the Ph.D. (and, if applicable, from the position held since the Ph.D.); particular attention is thus paid to French or international candidates who obtained their doctorate abroad.

Objectives of the Post-doctoral Position

This post-doctoral position aims at integrating the StarPU task-based runtime system as the scheduling and execution engine of the DEMI cardiac electrophysiology simulation code, to enable and optimise its execution on GPU-accelerated computing nodes. The original DEMI simulator follows a regular fork-join node-level execution scheme based on OpenMP parallel loops coupled to a MPI based distributed execution. Instead, task-based runtime systems such as StarPU take advantage of asynchronism to avoid idle times and overlap operations such as data transfers with computations. In order to maximally benefit from this asynchronism, care will have to be taken to express the entire flow of the application through data dependencies all along the simulation, without resorting to barriers. An important challenge of integrating StarPU as DEMI’s execution engine will thus be to delineate individual tasks carefully, with a sufficient computational grain to enable the efficient use of GPU accelerators, but small enough to ensure good load balancing, and to partition data structures inside DEMI to maximize parallelism and latency hiding in complex cardiac cells simulation layouts.

StarPU ships with a collection of scheduling algorithms with a variety of trade-offs between the scheduling capabilities and costs, to match distinct application characteristics. Another challenge will therefore be to select and fine-tune StarPU's scheduling algorithm to take into account, and take advantage of the potential specificities of cardiac electrophysiology simulations with DEMI, in terms of priorities, data access patterns and transfer volumes.
 

Principales activités

  • Bibliographical study on related works.
  • Port of DEMI on top of StarPU to target heterogeneous computing clusters of multi-GPU nodes.
  • Performance evaluation and tuning on synthetic meshes, on local computing clusters (PlaFRIM in Bordeaux).
  • Experimentations on national supercomputing resources such as GENCI and SLICES resources in France and the EX3 platform at Simula.
  • Publications on the findings of this work in selected conferences and journals.

 

Compétences

  • Language C programming on Linux systems
  • Parallel and distributed programming for HPC
  • Applied mathematics and scientific simulation programming
  • Development with Git / Gitlab / GitHub environments
  • Fluency in written and spoken English
 

Avantages

  • Subsidized meals
  • Partial reimbursement of public transport costs
  • Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
  • Possibility of teleworking and flexible organization of working hours
  • Professional equipment available (videoconferencing, loan of computer equipment, etc.)
  • Social, cultural and sports events and activities
  • Access to vocational training
  • Social security coverage

Rémunération

The monthly salary will be 2927€ (before social security contributions and monthly witholding tax).