Application-Driven Memory xTCO

Il y a 3 jours

VlaamsBrabant, Vlaams-Brabant, Belgique Imec Temps plein 11 000 € - 18 000 €/an

Currentmemory systems and hierarchies are struggling to keep pace with modern workloaddemands, especially those imposed by generative AI models. On-chip, SRAMscaling has stagnated due to physical scaling limits in advanced fabricationnodes. Meanwhile, off-chip, DRAM scaling is slowing due to challenges in layoutand connectivity. As a result, novel memory technologies must be consideredwholistically, both in terms of the technology proposition and theapplication-driven use case.

imec worksto identify and evaluate novel memory technologies, organizations,integrations, and packaging, including but not limited to string memories suchas charge-coupled device memories (CCD), flash memories, embedded DRAM, ferro-backedmemory, and others. On the Technology-Application Co-Optimization for Systems(TACOS) team, we specialize in end-to-end exploration of these memories andalternatives, i.e. from technology to application mapping, by usingsystem-technology co-optimization (STCO) methodologies.

Thisinternship proposes the exploration of such memories from a co-design point ofview, including both top-down (establishing technology requirements) andbottom‑up (establishing application paradigm shifts) approaches. During thisinternship the candidate will be responsible for,

  • Guided research of current and emerging memory technologies as well asinterfacing in-house experts.
  • Employing STCO methodology by way of using and connecting mixed‑fidelitymodelling and simulation tools. This can include analytical models,cycle‑aware/-approximate simulation (with tools such as Ramulator, DRAMsys, orgem5), and/or cycle accurate simulation.
  • Profiling and mapping applications with full‑and sub‑system awarenessfor technology propositions to quantify costs and trade‑offs of data movement.
  • Developingruntime optimisation strategies (e.g., adaptive tiling/scheduling andprefetching) to dynamically balance latency, bandwidth, and energy underchanging workload phases.

This role is ideal for someone who is deeplyinterested in hardware‑software‑technology codesign, memory architecture, andworking in an interdisciplinary environment that values innovation, creativity,and real‑world impact.

Profile: You are analytical and detail-oriented,with a strong interest in technology, subsystem-level modelling andhardware‑software codesign. You are adept at or have a keen interest in technologymodelling, performance and energy analysis tools, and cutting‑edge memoryresearch.

Type of internship: PhD internship, Master internship

Duration: 6‑12 months

Required educational background: Electromechanical engineering, Computer Science, Electrotechnics/Electrical Engineering

University promotor: Chris Van Hoof (KU Leuven)

Supervising scientist(s): For further information or for application, please contact Joshua Klein ( Joshua.Klein@imec.be )

The reference code for this position is 2026‑INT‑297. Mention this reference code in your application.

Imec allowance will be provided.

Applications should include the following information:

  • resume
  • motivation
  • current study

imec's cleanroom