Internship - Parallel algorithm for speculative execution in the task-based method

Le descriptif de l’offre ci-dessous est en Anglais

Type de contrat : Stage

Niveau de diplôme exigé : Bac + 4 ou équivalent

Autre diplôme apprécié : To be student in last year of a MS (niveau BAC+5)

Fonction : Stagiaire de la recherche

Contexte et atouts du poste

The Inria CAMUS/ICube ICPS research team focuses on parallelization, optimization, profiling, modeling, and compilation. The team has a growing interest in the approaches used and enhanced by the high-performance computing community.

In this context, the position involves the development of tools for creating high-performance applications, focusing on parallel algorithms as well as efficient and highly optimized implementations.

The results of this work will have direct applications in HPC and compilation.

This position is funded by the AUTOSPEC ANR JCJC project.

Mission confiée

The project aims to study speculative executions in the context of task-based parallelization. We described an initial model several years ago (https://peerj.com/articles/cs-183/). We now aim to extend it to support new types of uncertain accesses, and ideally create a fully generic system that works with all types of speculation.

To achieve this goal, it will be necessary first to understand the principles of task-based parallelization and speculative execution. Then, solutions will be designed conceptually before implementing them in practice.

We are also interested in evaluating the performance on large multicore systems.

Principales activités

We provide here a non-exhaustive list of possible key challenges:

  • Finding new algorithms to use speculative execution
  • Implementing these algorithms in our specx tool (https://gitlab.inria.fr/bramas/specx), or maybe first in a standalone PoC
  • Potentially adapt the core layer of specx to have an efficient concurrent system
  • Validate the tool by benchmarking on modern computing nodes, and continuously improve the tool or find out what are the potential limitations

 

Compétences

  • Student in Master degree (Computer Science or equivalent)
  • Proactive, high interest in solving problems
  • C++17 (and a bit of C)
  • Knowledge of parallelization is an asset

Languages :

  • English and/or French

 

The internship might be continued in a PhD but this will require obtaining a scholarship, which success will be tied to the marks and quality of the involved student.

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 (after 6 months of employment) 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