.png){kind=icon}
返回
HyperFPGA: A Scalable Reconfigurable Platform for Computing Paradigms Research
.png){kind=icon}
PRE
AI
DOI：10.1109/ACCESS.2025.3637300.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
In recent years, the limitations of von Neumann-based supercomputers in terms of energy efficiency and performance scalability have stressed the need for alternative computing paradigms. This paper introduces the HyperFPGA, an open SoC-FPGA cluster designed for experimental research on novel supercomputing architectures and paradigms. Unlike existing platforms, the HyperFPGA offers flexibility across a wide range of configurations, from electrical standards for data transmission to distributed computing models. Additionally, it is equipped with multiple power consumption monitors to independently measure different computational and hardware aspects, separating data movement from computation. The paper details the design and implementation of the HyperFPGA and demonstrates its basic functionality and scalability through a custom computational solution for the non-attacking n-queens problem, a well-known distributed computing problem. These preliminary tests demonstrate the ability of the HyperFPGA to efficiently handle computationally intensive and distributed tasks, showcasing its potential for fine-grained reconfigurable supercomputing solutions addressing a broad spectrum of scientific and engineering problems. The platform is available for open collaborative projects focused on energy-efficient supercomputing and computational physics.
Keyword:
Field programmable gate arrays
Hardware
Computer architecture
Power demand
Scalability
Protocols
Reconfigurable architectures
Monitoring
Standards
Software
FPGA
SoC
heterogeneous computing
supercomputing
reconfigurable computing
期刊
IF:
3.6
论文数:
9.8W
被引数:
29.4W
机构
引用论文
Formic: Cost-efficient and Scalable Prototyping of Manycore ArchitecturesFormic：多核架构的高性价比与可扩展原型设计
OAAI
OAAI分享收藏分享收藏
分享收藏分享收藏BiCoSS: Toward Large-Scale Cognition Brain With Multigranular Neuromorphic ArchitectureBiCoSS: 面向具有多颗粒神经形态结构的大规模认知大脑
PREAI
分享收藏Remote Laboratory for E-Learning of Systems on Chip and Their Applications to Nuclear and Scientific Instrumentation远程实验室用于片上系统及其实验电子学习的教学，及其在核科学与科学仪器中的应用.png)
Electronics
ElectronicsIF0
PREAI
分享收藏
查看更多内容