Dufy TEGUIA

Dufy TEGUIA

Currently PhD student at Université Grenoble Alpes in CIFRE with Orange Innovation. I like reading, abstract thinking, gaming and music

Posts by Collection

portfolio

publications

Analysis of a modern distributed hypervisor: what we learn from our experiments

Published in Systems for Post-Moore Architectures (SPMA), co-located with Eurosys 2022, Rennes, 2022

GiantVM is the state of the art distributed hypervisor (VEE 2020) which is based on KVM. It allows to start a guest OS whose CPU and memory are provided by several physical machines. In this paper, we study the origin of performance overhead of GiantVM – which is the DSM, hence, in term of number of page faults. Then we propose several optimizations which allow to improve performances by about 39%. Moreover, based on our study and proposed optimizations, we laid out guidelines to build a distributed hypervisor that is independent and flexible – easy to evolve, from a distributed shared memory (DSM) system.. Read more

Recommended citation: Mohamed Karaoui, Brice Teguia, Bernabe Batchakui,Alain Tchana Analysis of a modern distributed hypervisor: what we learn from our experiments, SPMA 2022 https://drive.google.com/file/d/1tVcswD44EEXhFoa0gMwYY87t5PHzHZxY/view?usp=sharing

vPIM – Virtualization of Processing-in-Memory

Published in Middleware 2024, 2024

Data movement is the leading cause of performance degradation and energy consumption in modern data centers. Processing in-memory (PIM) is an architecture that addresses data movement by bringing computation inside the memory chips. This paper is the first to study the virtualization of PIM devices by designing and implementing vPIM, an open-source UPMEM-based virtualization system for the cloud. Our vPIM design considers four requirements: Compatibility such that no hardware and no hypervisor changes are needed; Multiplexing and isolation for a higher utilization ratio; Utilizability and transparency such that applications written for PIM can be efficiently run out-of-the-box, leading to rapid adoption; Minimalization of virtualization performance overhead. We prototype vPIM in Firecracker, expanding the virtio standard. Our experimental evaluation uses 16 applications provided by PrIM, a recent PIM benchmark suite. The virtualization overhead is between 1.01× and 2.07× for untouched PrIM applications. To keep overhead low, vPIM introduces several optimizations: zero-copy from guest OS to Firecracker, efficient virtio queues management, efficient Guest Physical Address to Host Virtual Address translation, parallel processing on multiple ranks, automatic data batching and prefetching, and the reimplementation of some specific functionalities in C instead of Rust. We hope this work will lay the foundation for future research on PIM for cloud computing. Read more

Recommended citation: Dufy Teguia, Jiaxuan Chen, Stella Bitchebe, Oana Balmau, Alain Tchana, vPIM – Virtualization of Processing-in-Memory, Middleware 2024

talks