Travailler à TIMA

Propositions de stages


Self-adaptive approximate System on a Chip: Hardware platform development

Équipe : TIMA Laboratory - AMfoRS team - 46 avenue Félix Viallet - 38031 GRENOBLE Cedex

Date de début : January 2018

Durée : 6 months

Profil : The project SysAx proposed here aims at developing new methods and design techniques allowing the approximate System on a Chip (SoC) to self adapt to its environment. For example, the inputs may become noisy, the alert threshold of the supply battery may be reached, the chip may give signs of reliability degradation, ... The system must then switch from a precision mode to another to reduce its power consumption or to extend its lifetime. Existing and new HW/SW techniques will be gathered to constitute a self-adaptive platform running different RMS applications such as vision recognition and/or machine learning algorithms. The project will include two internships, one for the hardware development and one for the software development. This one is dedicated to the hardware part.

See complete information

Personne à contacter : Dr Mounir BENABDENBI (mounir.benabdenbi@univ-grenoble-alpes.fr)

Poursuite en thèse possible

Rémunération: 550 € net per month (regular internship "gratification")

Niveau: Master

 

 

Self-adaptive approximate System on a Chip: Software platform development

Équipe : TIMA Laboratory - AMfoRS team - 46 avenue Félix Viallet - 38031 GRENOBLE Cedex

Date de début : January 2018

Durée : 6 months

Profil : The project SysAx proposed here aims at developing new methods and design techniques allowing the approximate System on a Chip (SoC) to self adapt to its environment. For example, the inputs may become noisy, the alert threshold of the supply battery may be reached, the chip may give signs of reliability degradation, ... The system must then switch from a precision mode to another to reduce its power consumption or to extend its lifetime. Existing and new HW/SW techniques will be gathered to constitute a self-adaptive platform running different RMS applications such as vision recognition and/or machine learning algorithms. The project will include two internships, one for the hardware development and one for the software development. This one is dedicated to the software part.

See complete information

Personne à contacter : Dr Mounir BENABDENBI (mounir.benabdenbi@univ-grenoble-alpes.fr)

Poursuite en thèse possible

Rémunération: 550 € net per month (regular internship "gratification")

Niveau: Master