Network Solution for Exascale Architectures

Livrables

Optimized MPI and compute in network implementation

Report the final optimizations on MPI and on in-network compute prototype.NB: ParTec, the linked third party of FZJ, will lead D4.7.

Planned MPI-related optimizations

Report the first designs and specifications of T45

Communication and Dissemination final report

This report will contain a summary of communication and dissemination activities, featuring the metrics defined in D5.1.

Final report [all tasks]

Description of the technical and scientific results of the project.

Communication and Dissemination intermediate report

This report will contain a summary of communication and dissemination activities, featuring the metrics defined in D5.1.

RoCE and IPoverBXI evaluation report

This document produced by tasks 21 and 22 will include the results of the experiments and the findings for RoCE and IPoverBXI communications

Collaboration plan

Collaboration plan with definition of common objectives and activities including milestonesThe updated versions will be provided at the end of each reporting period

Report on holistic evaluation of RED-SEA network technologies

It will report the key technologies developed in RED-SEA the findings that support them and their implications for extreme-scale systems and for EPI processors.

Public website

The public website will be considered in this deliverable It will subsequently be maintained and updated to reflect the projects life

BXI to Ethernet bridging demonstrator

This deliverable produced by task 2.6 is the Ethernet Gateway testbed and its validation report.

Publications

rFaaS: Enabling High Performance Serverless with RDMA and Decentralization

Auteurs: Marcin Copik, Konstantin Taranov, Alexandru Calotoiu, Torsten Hoefler
Publié dans: 2021
Éditeur: Arxiv
DOI: 10.48550/arxiv.2106.13859

FaaSKeeper: Learning from Building Serverless Services with ZooKeeper as an Example

Auteurs: Marcin Copik, Alexandru Calotoiu, Konstantin Taranov, Torsten Hoefler
Publié dans: 2023
Éditeur: Cornell University
DOI: 10.48550/arxiv.2203.14859

Cppless: Productive and Performant Serverless Programming in C++

Auteurs: Lukas Möller, Marcin Copik, Alexandru Calotoiu, Torsten Hoefler
Publié dans: 2024
Éditeur: Cornell University
DOI: 10.48550/arxiv.2401.10834

Optimized Page Fault Handling During RDMA

Auteurs: Antonis Psistakis; Nikos Chrysos; Fabien Chaix; Marios Asiminakis; Michalis Gianioudis; Pantelis Xirouchakis; Vassilis Papaefstathiou; Manolis Katevenis
Publié dans: IEEE Transactions on Parallel and Distributed Systems, Numéro Vol 33, Numéro 12, 2022, Page(s) 3990-4005, ISSN 1045-9219
Éditeur: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tpds.2022.3175666

Canary: Congestion-aware in-network allreduce using dynamic trees

Auteurs: Daniele De Sensi; Edgar Costa Molero; Salvatore Di Girolamo; Laurent Vanbever; Torsten Hoefler
Publié dans: Future Generation Computer Systems, 152, Numéro 152, 2023, Page(s) 70-82, ISSN 1872-7115
Éditeur: Elsevier
DOI: 10.1016/j.future.2023.10.010

Influence of Network Performance Variability on Application Scalability

Auteurs: Daniele De Sensi; Tiziano De Matteis; Konstantin Taranov; Salvatore Di Girolamo; Tobias Rahn; Torsten Hoefler
Publié dans: Proceedings of the ACM on Measurement and Analysis of Computing Systems, 6 (3), Numéro 1, 2022, Page(s) 1-27, ISSN 2476-1249
Éditeur: ACM
DOI: 10.1145/3570609

Asynchronous Distributed-Memory Triangle Counting and LCC with RMA Caching

Auteurs: Andras Strausz; Flavio Vella; Salvatore Di Girolamo; Maciej Besta; Torsten Hoefler
Publié dans: 2022 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Numéro 1, 2022, Page(s) 291-301
Éditeur: IEEE
DOI: 10.1109/ipdps53621.2022.00036

Lifting C Semantics for Dataflow Optimization

Auteurs: Calotoiu, Alexandru; Ben-Nun, Tal; Kwasniewski, Grzegorz; de Fine Licht, Johannes; id_orcid0000-0002-1500-7411; Schneider, Timo; Schaad, Philipp; Hoefler, Torsten
Publié dans: ICS '22: Proceedings of the 36th ACM International Conference on Supercomputing, Numéro 1, 2022, ISBN 978-1-4503-9281-5
Éditeur: ACM
DOI: 10.48550/arxiv.2112.11879

User-guided Page Merging for Memory Deduplication in Serverless Systems

Auteurs: Qiu, Wei; Copik, Marcin; Wang, Yun; Calotoiu, Alexandru; Hoefler, Torsten
Publié dans: 2023 IEEE International Conference on Big Data (Big Data), Numéro 1, 2023, ISBN 979-8-3503-2445-7
Éditeur: IEEE
DOI: 10.1109/bigdata59044.2023.10386487

KafkaDirect: Zero-copy Data Access for Apache Kafka over RDMA Networks

Auteurs: Taranov, Konstantin; Byan, Steve; Hoefler, Torsten; Hoefler, Torsten
Publié dans: SIGMOD '22: Proceedings of the 2022 International Conference on Management of Data, Numéro 1, 2022, ISBN 9781450392495
Éditeur: ACM
DOI: 10.1145/3514221.3526056

HEAR: Homomorphically Encrypted Allreduce

Auteurs: Marcin Chrapek, Mikhail Khalilov, Torsten Hoefler
Publié dans: SC '23: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2023, Page(s) 1-17
Éditeur: ACM
DOI: 10.1145/3581784.3607099

Proyecto RED-SEA: Resultados Intermedios

Auteurs: Duro, José; Castello, Adrian; Gomez, Maria E.; Sahuquillo, Julio; Quintana, Enrique; Gomez, Gabriel; Sanchez, Miguel; Escudero-Sahuquillo, Jesus; Garcia, Pedro J.; Alfaro, Francisco J.; Sanchez, José L.; Quiles, Francisco J.
Publié dans: Proceedings of the XXXIII Jornadas de Paralelismo (JP23), 2023, ISBN 978-84-09-54466-0
Éditeur: Zenodo
DOI: 10.5281/zenodo.10688152

FMI: Fast and Cheap Message Passing for Serverless Functions

Auteurs: Marcin Copik; Roman Böhringer; Alexandru Calotoiu; Torsten Hoefler
Publié dans: ICS'23: Proceedings of the 37th International Conference on Supercomputing, Numéro 1, 2023, Page(s) 373-385
Éditeur: ACM
DOI: 10.1145/3577193.3593718

flexible in-network allreduce

Auteurs: Daniele De Sensi; Salvatore Di Girolamo; Saleh Ashkboos; Shigang Li; Torsten Hoefler
Publié dans: SC '21: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, Numéro 1, 2021
Éditeur: ACM
DOI: 10.1145/3458817.3476178

RED-SEA: Network Solution for Exascale Architectures

Auteurs: Andrea Biagioni; Paolo Cretaro; Ottorino Frezza; Francesca Lo Cicero; Alessandro Lonardo; Michele Martinelli; Pier Stanislao Paolucci; Elena Pastorelli; Francesco Simula; Matteo Turisini; Piero Vicini; Roberto Ammendola; Pascale Bernier-Bruna; Claire Chen; Said Derradji; Stephane Guez; Pierre-Axel Lagadec; Gregoire Pichon; Etienne Walter; Gaetan De Gassowski; Matthieu Hautreaux; Stephane Mathieu;
Publié dans: 2022 25th Euromicro Conference on Digital System Design (DSD), Numéro 1, 2022
Éditeur: IEEE
DOI: 10.1109/dsd57027.2022.00100

Building blocks for network-accelerated distributed file systems

Auteurs: Di Girolamo, Salvatore; De Sensi, Daniele; Taranov, Konstantin; Malesevic, Milos; Besta, Maciej; Schneider, Timo; Kistler, Severin; Hoefler, Torsten
Publié dans: SC '22: Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis, 2022, Page(s) 1-14
Éditeur: IEEE Press
DOI: 10.5555/3571885.3571898

NeVerMore: Exploiting RDMA Mistakes in NVMe-oF Storage Applications

Auteurs: Taranov, Konstantin; Rothenberger, Benjamin; De Sensi, Daniele; Perrig, Adrian; Hoefler, Torsten
Publié dans: CCS '22: Proceedings of the 2022 ACM SIGSAC Conference on Computer and Communications Security, 2022, Page(s) 2765-2778, ISBN 978-1-4503-9450-5
Éditeur: ACM
DOI: 10.1145/3548606.3560568

A RISC-V in-network accelerator for flexible high-performance low-power packet processing

Auteurs: Salvatore Di Girolamo; Andreas Kurth; Alexandru Calotoiu; Thomas Benz; Timo Schneider; Jakub Beranek; Luca Benini; Torsten Hoefler
Publié dans: 2021 ACM/IEEE 48th Annual International Symposium on Computer Architecture (ISCA), Numéro 2, 2021
Éditeur: IEEE
DOI: 10.1109/isca52012.2021.00079

A RDMA Interface for Ultra-Fast Ultrasound Data-Streaming over an Optical Link

Auteurs: Andrea Cossettini; Konstantin Taranov; Christian Vogt; Michele Magno; Torsten Hoefler; Luca Benini
Publié dans: Proceedings of the 2022 Design, Automation & Test in Europe Conference & Exhibition (DATE), 2022, ISBN 978-3-9819263-6-1
Éditeur: IEEE
DOI: 10.23919/date54114.2022.9774599

Livrables

Publications

Partager cette page

Télécharger