Research
The Parallel and Distributed Systems group is currently involved in the following research projects.
Rackscale computing
- Scalevisor (ANR project 2019-2022) is an hypervisor at the scale of a rack.
Performance analysis for parallel applications
- EZTrace is a tracing framework for HPC applications.
- NumaMMA is a memory profiler.
- Pythia (ANR project 2019-2022) is a profiler and a runtime that optimizes applications by predicting their behavior.
Energy consumption for HPC applications
- DUF is a runtime systems that automatically adapts the processor uncore frequency in order to save energy with as little performance degradation as possible.
Scalable File Systems
- RainbowFS ANR project (2017-2020) that investigates an approach to distributed storage that ensures consistency semantics tailored to the application, while retaining scalability and availability.
Runtime for HPC and IA applications
- This project has the goal of designing new runtimes and optimizing existing runtimes in order to execute more efficiently HPC and IA applications.
Serverless Computing
- CloudButton is an H2020 project (2019-2021) that aims at creating a Serverless Data Analytics Platform.
Multiscale Distributed Event-Based Systems
Privacy
- Primate is a ANR project (2018-2022) that has the goal of proposing a domain specific language for C and C++ applications tailored to enforce privacy.
