I am currently involved in the following projects and activities:

      (PI, Technical Coordinator, EU Horizon 2020, 780245)

      In this project we will research and develop end-to-end solutions for dynamic and automatic

      exploitation of heterogeneous hardware on cloud-deployed Big Data stacks.

       (Co-I, EU Horizon 2020, 732366)

In this project we aim to develop hyperscale JVMs on top of aggregated cloud resources.

We plan to extend memory allocation and GC algorithms in NUMA settings able to handle TBs of memory.

In this project we focus on low-power heterogeneous execution of JVMs.

We mainly target demanding portable Computer Vision applications implemented in Java and other programming languages. Heterogeneous execution (GPUs, FPGAs, etc.), hw/sw co-designed optimizations,

scheduling, etc., on ARM and x86 architectures are explored.

Beehive aims to provide a state-of-the-art hw/sw co-designed stack that will enable researchers

to perform vertical research (from the application code down to the hardware; real or simulated).

We target a variety of programming languages, runtimes and architectures.

Please checkout Beehive's github pages for further information about the individual components here.

Older projects

© 2020 by Christos Kotselidis.