Master projects/internships - Leuven | More than two weeks ago
Identifying HPC-class targets for emerging compute-near-memory architectures.
Systems that deploy computational logic near memory can overcome typical von Neumann-based bottlenecks (e.g. memory wall) by limiting the amount of data transferred to central compute areas in a system. With modern compute-near memory (CnM) systems still in their infancy and typical programming paradigms focused on centralized computing however, new work must be undertaken to identify, implement, and evaluate new software for CnM hardware.
During this research internship, you will find distributed computation targets by becoming acquainted with modern datacentre-class algorithms and applications, learning about application characterization, and using state-of-the-art methodologies and tools to identify memory-based bottlenecks. Your work will build the foundation for a framework with which modern “typical” applications can become CnM-enabled applications and will be vital for guiding CnM hardware development.
Key responsibilities will include:
Profile: You are analytical and detail-oriented, with a strong interest in AI, genomics, databases, forecasting, or other datacentre-class algorithms. You are adept at or have a keen interest in programming, static and dynamic analysis, and performance evaluation tools.
Background: Currently pursuing or already have a degree in computer engineering, computer science, informatics, or electrical engineering. Has some background in algorithms, parallel programming, or accelerators. Knowledge of object-oriented programming and scripting languages is an advantage.
Type of project: Internship, Thesis
Required degree: Master of Science, Master of Engineering Science, Master of Engineering Technology
Required background: Computer Science, Electrotechnics/ Electrical Engineering
Duration: 3 to 9 months
Supervising scientist(s): For more information on this topic, please contact Leandro M. Giacomini Rocha (leandro.m.giacominirocha@imec.be) and Joshua Klein (joshua.klein@imec.be).