OPEN AND PROGRAMMABLE ACCELERATORS FOR DATA-INTENSIVE APPLICATIONS IN THE CLOUD

Informations projet

CHORYS

N° de convention de subvention: 101189551

DOI

10.3030/101189551

Date de signature de la CE 15 Novembre 2024

Date de début 1 Janvier 2025

Date de fin 31 Decembre 2028

Financé au titre de

• Digital, Industry and Space

Coût total

€ 4 776 602,50

Contribution de l’UE

€ 4 776 602,50

Coordonné par

KOBENHAVNS UNIVERSITET
Denmark

CORDIS fournit des liens vers les livrables publics et les publications des projets HORIZON.

Les liens vers les livrables et les publications des projets du 7e PC, ainsi que les liens vers certains types de résultats spécifiques tels que les jeux de données et les logiciels, sont récupérés dynamiquement sur OpenAIRE .

Livrables

Data Management Plan (1)

Data Management Plan

The data management plan will describe the characteristics of input and measurement data as well as software artefacts according to FAIR data principles.

Documents, reports (10)

Relevant Accelerator-Based Cloud Environment

Definition of the relevant accelerator-based cloud environment to be used for the three demonstrators.

Refined Exploitation Plans

Refined exploitation plans and business cases for the project partners.

Design Space for RISC-V extensions

Design Space for RISC-V extensions well-suited for near-data processing with data-intensive patterns.

Project and Risk Management Plan

Definition of the project management processes and tools in the CHORYS project, including a risk management plan supported by monitoring and reporting tools.

Requirements and Gap Analysis

Specification of data-intensive data patterns, definition of gaps in terms of component and system integration of an accelerator-based cloud stack built with RISC-V cores, and definition of relevant metrics for the project.

Initial use cases

Definition of three use cases based on CYSO requirements and based on use case workshop with stakeholders.

Design Space and Exploration Methodology

Definition of the design space and methodology for component-based integration of RISC-V cores on FPGA boards for co-designed SSD and NIC.

Refined System Architecture

Refinement of the CHORYS system architecture based on insight gained in the technical work packages.

Stakeholder Engagement Plan

Plan for workshops, consultations and identification of relevant international events

Initial System Architecture

Definition of the CHORYS system architecture for an accelerator-based cloud computing hardware and software stack co-designed for data-intensive workloads

Demonstrators, pilots, prototypes (3)

Library of Host-Based Data-Intensive Patterns

Library of data-intensive data patterns in data parallel C++, executing on a host with data loaded in memory from SSD or network.

Initial OpenNIC-V Prototype

Prototype of an FPGA-based smart NIC equipped with RISC-V cores.

Initial FPGA-Based Platforms for Co-Designed SSDs and NICs

Initial design and implementation of a FPGA-based platform that integrates RISC-V cores with memory and specialized IP for near-data processing with co-designed SSDs and NICs.

Websites, patent fillings, videos etc. (1)

Communication Toolkit

Logo, Documents template, Project content collaboration platform, Project website, Project presence on Linkedin

Publications

Conference proceedings (11)

What No One Tells You About Page Pruning in Parquet: The Real Cost of Page Index Parsing

Auteurs: Faeze Faghih, Si Jun Kwon, Zsolt István
Publié dans: Proceedings of the 1st International Workshop on Data FORMATS for Modern Architectures and Workloads, 2026, ISSN 979-8-4007-2635
Éditeur: ACM
DOI: 10.1145/3802514.38091

Model-Driven Right-Sizing of Offloading in Data Processing Pipelines

Auteurs: Faeze Faghih, Maximilian Hüttner, Florin Dinu, Zsolt István
Publié dans: Proceedings of the 21st International Workshop on Data Management on New Hardware, 2025, ISSN 979-8-4007-1940
Éditeur: ACM
DOI: 10.1145/3736227.3736239

Enabling Fast Networking in the Public Cloud

Auteurs: Alireza Sanaee, Vahab Jabrayilov, Ilias Marinos, Farbod Shahinfar, Divyanshu Saxena, Gianni Antichi, Kostis Kaffes
Publié dans: Proceedings of the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2, 2026
Éditeur: ACM
DOI: 10.1145/3779212.3790158

Don't Stall Me Now: Hiding Memory Latency in eBPF

Auteurs: Farbod Shahinfar, Marco Molè, Aurojit Panda, Gianni Antichi
Publié dans: Proceedings of the ACM SIGCOMM 2026 Conference, 2026
Éditeur: ACM
DOI: 10.1145/3779212.3790158

Demystifying Performance of eBPF Network Applications

Auteurs: Farbod Shahinfar, Sebastiano Miano, Aurojit Panda, Gianni Antichi
Publié dans: Proceedings of the ACM on Networking, Numéro 3, 2025, ISSN 2834-5509
Éditeur: Association for Computing Machinery (ACM)
DOI: 10.1145/3749216

SSD-iq: Uncovering the Hidden Side of SSD Performance

Auteurs: Gabriel Haas, Bohyun Lee, Philippe Bonnet, Viktor Leis
Publié dans: Proceedings of the VLDB Endowment, Numéro 18, 2025, ISSN 2150-8097
Éditeur: Association for Computing Machinery (ACM)
DOI: 10.14778/3749646.3749694

De²Dup: Extended Deduplication for Multi-Tenant Databases

Auteurs: Alexander Krause, Jannis Kowalick, Johannes Pietrzyk, Dirk Habich, Wolfgang Lehner
Publié dans: Proceedings of the 21st International Workshop on Data Management on New Hardware, 2025
Éditeur: ACM
DOI: 10.1145/3736227.3736236

Rhea: a Framework for Fast Design and Validation of RTL Cache-Coherent Memory Subsystems

Auteurs: Davide Zoni, Andrea Galimberti, Adriano Guarisco
Publié dans: 2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD), 2025
Éditeur: IEEE
DOI: 10.1109/ICCAD66269.2025.11240659

Performance Implications at the Intersection of AF_XDP and Programmable NICs

Auteurs: Marco Molè, Farbod Shahinfar, Francesco Maria Tranquillo, Davide Zoni, Aurojit Panda, Gianni Antichi
Publié dans: Proceedings of the 3rd Workshop on eBPF and Kernel Extensions, 2025
Éditeur: ACM
DOI: 10.1145/3748355.3748359

Steering Memory Deduplication in Multi-Tenant Database Systems

Auteurs: Jannis Kowalick, Alexander Krause, Dirk Habich, Wolfgang Lehner
Publié dans: Proceedings of the 36th GI-Workshop Grundlagen von Datenbanken, 2025
Éditeur: CEUR Workshop Proceedings

S2VEC: Compiler-Driven Stream Specialization for Linearized Vectorization

Auteurs: Luís Crespo, Ana Fernandes, Gabriel Falcao, Pedro Tomás, Nuno Roma, Nuno Neves
Publié dans: Proceedings of the 40th ACM International Conference on Supercomputing, 2026
Éditeur: ACM
DOI: 10.1145/3797905.3807866