Heuristics for Heterogeneous Memory (H2M)

Participant Institutions: Inria (TADaaM team), RWTH Aechen

Organizing Institution: Agence Nationale de la Recherche (ANL)

Type: Call to projects

Participation Year: 2021 - Currently

Role: PhD Researcher

Leaders: Brice Goglin, Christian Terboven

Research Area: Heterogeneous Memory


High-performance computing (HPC) is crucial to advance computational science and engineering. In all modern computing systems, the performance gap between compute and memory continues to spread, particularly in the face of multi-core and accelerated systems. In consequence, the memory subsystem is changing: the evolution of the cache hierarchy is followed by new technologies with new kinds of memory. In the context of HPC, this has been pioneered by combining traditional main memory with a small fraction of high-bandwidth memory. In systems with accelerators, like GPUs, the heterogeneity by means of different kinds of memory is already higher. Currently, applications have to be heavily modified for specific target platforms, and have to employ vendor-specific APIs to exploit heterogeneous memory.

H2M combines the Inria team’s expertise in exposing low-level runtime functionality and the RWTH’s group ability to leverage these to develop abstractions for HPC programming. The result of this joint work will be a deep understanding of how heterogeneous memory systems have to be programmed, a hierarchy of programming abstractions and a set of heuristics for use in intelligent runtime systems to serve applications optimized for performance and scalability.


Real-time Middleware Based on the Publish / Subscribe Model

Participant Institutions: EPN, UPS, UTPL

Organizing Institution: CEDIA


Participation Year: Sep/2014-Aug/2015

Role: Research Technical Assistant

Leaders: Ernesto Jiménez Merino, Xavier Calderón, Bertha Tacuri, Samantha Cueva, Patricia Ludeña

Research Area: Information and communication technologies, Distributed systems

Status: Finished


The objective of this project is to develop a middleware for distributed systems based on events (or publication / subscription) that serves as the basis for researching advanced techniques. Using DDS / RTPS ideas as a technological base, an open and flexible implementation will be developed that allows exploring the practical application of techniques based on reflection, AOP, dynamic generation of proxies and interception of calls. The objective of this is to incorporate as many of these advanced techniques as possible to our middleware. Similarly, this prototype will allow us to have the basis to explore other relevant aspects such as fault tolerance and performance on a very large scale in new projects.