Experimenting with Emerging RISC-V Systems for Decentralised Machine Learning

被引:5
|
作者
Mittone, Gianluca [1 ]
Tonci, Nicolo [2 ]
Birke, Robert [1 ]
Colonnelli, Iacopo [1 ]
Medic, Doriana [1 ]
Bartolini, Andrea [3 ]
Esposito, Roberto [1 ]
Parisi, Emanuele [3 ]
Beneventi, Francesco [3 ]
Polato, Mirko [1 ]
Torquati, Massimo [2 ]
Benini, Luca [3 ]
Aldinucci, Marco [1 ]
机构
[1] Univ Turin, Turin, Italy
[2] Univ Pisa, Pisa, Italy
[3] Univ Bologna, Bologna, Italy
关键词
Federated Learning; Edge Computing; RISC-V; Energy Consumption; Green Computing;
D O I
10.1145/3587135.3592211
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Decentralised Machine Learning (DML) enables collaborative machine learning without centralised input data. Federated Learning (FL) and Edge Inference are examples of DML. While tools for DML (especially FL) are starting to flourish, many are not flexible and portable enough to experiment with novel processors (e.g., RISC-V), non-fully connected network topologies, and asynchronous collaboration schemes. We overcome these limitations via a domain-specific language allowing us to map DML schemes to an underlying middleware, i.e. the FastFlow parallel programming library. We experiment with it by generating different working DML schemes on x86-64 and ARM platforms and an emerging RISC-V one. We characterise the performance and energy efficiency of the presented schemes and systems. As a byproduct, we introduce a RISC-V porting of the PyTorch framework, the first publicly available to our knowledge.
引用
收藏
页码:73 / 83
页数:11
相关论文
共 50 条
  • [21] Information Flow Tracking in RISC-V
    Shirley, Geraldine
    Saqib, Fareena
    2019 IEEE 16TH INTERNATIONAL CONFERENCE ON SMART CITIES: IMPROVING QUALITY OF LIFE USING ICT, IOT AND AI (IEEE HONET-ICT 2019), 2019, : 199 - 200
  • [22] RISC-V RANDOM TEST GENERATOR
    Dai Duong Tran
    Thi Giang Truong
    Truong Giang Do
    The Duc Do
    2021 15TH INTERNATIONAL CONFERENCE ON ADVANCED COMPUTING AND APPLICATIONS (ACOMP 2021), 2021, : 150 - 155
  • [23] A Framework for Fault Tolerance in RISC-V
    Doerflinger, Alexander
    Kleinbeck, Benedikt
    Albers, Mark
    Michalik, Harald
    Moya, Martin
    2022 IEEE INTL CONF ON DEPENDABLE, AUTONOMIC AND SECURE COMPUTING, INTL CONF ON PERVASIVE INTELLIGENCE AND COMPUTING, INTL CONF ON CLOUD AND BIG DATA COMPUTING, INTL CONF ON CYBER SCIENCE AND TECHNOLOGY CONGRESS (DASC/PICOM/CBDCOM/CYBERSCITECH), 2022, : 163 - 170
  • [24] Efficient Cryptography on the RISC-V Architecture
    Stoffelen, Ko
    PROGRESS IN CRYPTOLOGY - LATINCRYPT 2019, 2019, 11774 : 323 - 340
  • [25] Towards a firmware TPM on RISC-V
    Boubakri, Marouene
    Chiatante, Fausto
    Zouari, Belhassen
    PROCEEDINGS OF THE 2021 DESIGN, AUTOMATION & TEST IN EUROPE CONFERENCE & EXHIBITION (DATE 2021), 2021, : 647 - 650
  • [26] Maxpool operator for RISC-V processor
    Nevezi-Strango, David
    Rotar, Danut
    Valcan, Sorin
    Gaianu, Mihail
    2023 25TH INTERNATIONAL SYMPOSIUM ON SYMBOLIC AND NUMERIC ALGORITHMS FOR SCIENTIFIC COMPUTING, SYNASC 2023, 2023, : 246 - 250
  • [27] Is Open Hardware Worthwhile? Learning from Thales' Experience with RISC-V
    Legenvre, Herve
    Kauttu, Pietari
    Bos, Martin
    Khawand, Roger
    RESEARCH-TECHNOLOGY MANAGEMENT, 2020, 63 (04) : 44 - 53
  • [28] RISC-V的“芯”技术
    俞灵琦
    华东科技, 2021, (02) : 18 - 21
  • [29] RISC-V3: A RISC-V Compatible CPU With a Data Path Based on Redundant Number Systems
    Reichenbach, Marc
    Knoedtel, Johannes
    Rachuj, Sebastian
    Fey, Dietmar
    IEEE ACCESS, 2021, 9 : 43684 - 43700
  • [30] A Low Power Branch Prediction for Deep Learning on RISC-V Processor
    Sun, Mingjian
    Li, Yuan
    Chen, Song
    Kang, Yi
    2021 IEEE 32ND INTERNATIONAL CONFERENCE ON APPLICATION-SPECIFIC SYSTEMS, ARCHITECTURES AND PROCESSORS (ASAP 2021), 2021, : 203 - 206