CIFRE PhD Thesis«Architectural Exploration&Optimization on Complex SoC
Continental
Toulouse, France
il y a 15h

Description

CIFRE PhD Thesis Architectural Exploration and Optimization on Complex SoCs (en)

At Continental, we’re working to make mobility clean, safe, and intelligent. We consider digitalization as a strong enabler to further improve our contribution to these 3 goals.

Digitalization is not only transforming the way automotive systems are built and used, it will disrupt the whole people mobility experience.

Continental Digital Services France (CDSF) is a new subsidiary of Continental Automotive France that has been created to address these digitalization opportunities around mobility services and autonomous driving.

We want to achieve a fusion between on-board intelligence and the one of our platform in the cloud .

Our cloud assistant , dedicated to every connected car, can have real-time access to information far beyond the horizon of its on-board sensors.

At a much larger scale, it enables both holistic and historical analyses of the flow of all vehicles, while preserving the privacy of connected cars users.

Modern SoCs such as ACAP of Xilinx or DAHLIA from the space domain offer today : a high level of hardware parallelism, a great architectural diversity combining various architectures on the same SoC, a great diversity for determinism level, a great diversity of implementation models and a great diversity of computation models.

Exploit platforms in an effective and optimal, let’s say efficient, manner that present this level of diversity is an extreme complex task because it requires to combine multiple very specific and mostly sparse competences.

It requires to take into account very diverse constraints. It uses and need very diverse tools and technologies. It implies to make architectural choices all along the development cycle.

Making architectural choices requires methods and technical means (modelisation tools, simulation tools, benchmarks, etc.

Today there are no methods or tools covering the overall problem statement mentionned above, but several software bricks exist that can be used.

Il n’existe pas de méthodes ou moyens couvrant l’ensemble des problématiques évoquées, mais il existe de nombreuses briques logicielles sur lesquelles s’appuyer.

Some of them are currently being used.

In this PhD thesis, the work will be to define a method and the tool chain to permit architectural hardware and software exploration on computing architectures, heterogeneous multicore SoC type combining several cores of different types, SIMD units, vectorial accelerating units, FPGAs, GPUs and AI units.

Job description :

  • Thesis work (3 years)
  • Thesis work is related to the last paragraph of the introduction just over
  • Participation to the work of Continental teams
  • Time shared between the Continental site in Toulouse and the academic lab
  • Thesis co-funded by ANRT
  • Candidate profile :

  • Engineer or Master 2 in Applied Mathematics with skills in modeling
  • Good programming skills (C / C++, Java, Python)
  • Good knowledge in Artificial Intelligence and Machine Learning
  • Knowledge in mathematics modeling
  • Good level in English
  • Dynamic, you can take initiatives, and you are a good teammate Qualifications CIFRE PhD Thesis Architectural Exploration and Optimization on Complex SoCs (en)
  • At Continental, we’re working to make mobility clean, safe, and intelligent. We consider digitalization as a strong enabler to further improve our contribution to these 3 goals.

    Digitalization is not only transforming the way automotive systems are built and used, it will disrupt the whole people mobility experience.

    Continental Digital Services France (CDSF) is a new subsidiary of Continental Automotive France that has been created to address these digitalization opportunities around mobility services and autonomous driving.

    We want to achieve a fusion between on-board intelligence and the one of our platform in the cloud .

    Our cloud assistant , dedicated to every connected car, can have real-time access to information far beyond the horizon of its on-board sensors.

    At a much larger scale, it enables both holistic and historical analyses of the flow of all vehicles, while preserving the privacy of connected cars users.

    Modern SoCs such as ACAP of Xilinx or DAHLIA from the space domain offer today : a high level of hardware parallelism, a great architectural diversity combining various architectures on the same SoC, a great diversity for determinism level, a great diversity of implementation models and a great diversity of computation models.

    Exploit platforms in an effective and optimal, let’s say efficient, manner that present this level of diversity is an extreme complex task because it requires to combine multiple very specific and mostly sparse competences.

    It requires to take into account very diverse constraints. It uses and need very diverse tools and technologies. It implies to make architectural choices all along the development cycle.

    Making architectural choices requires methods and technical means (modelisation tools, simulation tools, benchmarks, etc.

    Today there are no methods or tools covering the overall problem statement mentionned above, but several software bricks exist that can be used.

    Il n’existe pas de méthodes ou moyens couvrant l’ensemble des problématiques évoquées, mais il existe de nombreuses briques logicielles sur lesquelles s’appuyer.

    Some of them are currently being used.

    In this PhD thesis, the work will be to define a method and the tool chain to permit architectural hardware and software exploration on computing architectures, heterogeneous multicore SoC type combining several cores of different types, SIMD units, vectorial accelerating units, FPGAs, GPUs and AI units.

    Job description :

  • Thesis work (3 years)
  • Thesis work is related to the last paragraph of the introduction just over
  • Participation to the work of Continental teams
  • Time shared between the Continental site in Toulouse and the academic lab
  • Thesis co-funded by ANRT
  • Candidate profile :

  • Engineer or Master 2 in Applied Mathematics with skills in modeling
  • Good programming skills (C / C++, Java, Python)
  • Good knowledge in Artificial Intelligence and Machine Learning
  • Knowledge in mathematics modeling
  • Good level in English
  • Dynamic, you can take initiatives, and you are a good teammate
  • Signaler cette offre d'emploi
    checkmark

    Thank you for reporting this job!

    Your feedback will help us improve the quality of our services.

    Postuler
    Mon email
    En cliquant sur « Continuer », je consens au traitement de mes données et à recevoir des alertes email, tel que détaillé dans la Politique de confidentialité de neuvoo. Je peux retirer mon consentement ou me désinscrire à tout moment.
    Continuer
    Formulaire de candidature