CRI: CI-New: OpenPiton 2: Enabling Open Source Manycore Hardware Research

计算机体系结构的研究一直受到实验基础设施薄弱的阻碍。特别是，在支持跨计算堆栈的多个层的研究（诸如硬件、基本输入输出系统（BIOS）、管理程序、操作系统、驱动程序、应用程序）、需要在新硬件上高速运行的复杂软件堆栈的全系统研究、依赖于真实的硬件实现的研究以及大规模研究方面，基础设施一直缺乏。OpenPiton 2是专门针对学术界的开源处理器基础设施，使研究人员能够使用真实的硬件设计而不仅仅是模拟器来完成大规模研究。OpenPiton 2项目创建了一个学术基础设施，允许研究人员设计和构建大规模的众核处理器。特别是，这项研究调查了不同的输入/输出类型，使研究人员能够在不同的FPGA制造商的FPGA上使用OpenPiton基础设施，并为学术研究人员提供了一个托管的FPGA平台，以运行高核心数的OpenPiton 2设计，这是典型的个别大学资源所无法实现的。这项研究的重点是支持学术界，因此具有很大的学术推广组成部分。这项工作可以对科学界和更广泛的世界产生巨大的影响，使更有效的微处理器设计，从而节省能源，使更多的科学研究得以完成。这个奖项反映了NSF的法定使命，并已被认为是值得支持的评估使用基金会的智力价值和更广泛的影响审查标准。

Computer Architecture research has been traditionally hampered by poor experimental infrastructure. In particular, infrastructure has been lacking in terms of supporting research that straddles multiple layers of the computing stack such as hardware, Basic Input-Output System (BIOS), hypervisor, operating system, drivers, applications, full system research that necessitates complex software stacks running at speed on novel hardware, research that depends on real hardware implementation, and research at scale. OpenPiton 2 is an open source processor infrastructure specifically targeted at the academic community that enables researchers to complete research at scale using real hardware designs and not simply simulators. The OpenPiton 2 project creates an academic infrastructure that allows researchers to design and build many-core processors of large scale. In particular, this research investigates different Input/Output types, enables researchers to use the OpenPiton infrastructure on different FPGA manufacture's FPGAs and provides a hosted FPGA platform to run high core-count OpenPiton 2 designs for academic researchers that are otherwise not possible with typical individual university resources. This research focuses on supporting the academic community and hence features a large academic outreach component. This work can have large impacts on the scientific community and broader world by enabling more efficient microprocessors to be designed thereby saving energy and enabling more scientific research to be accomplished.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.