Non-deterministic event brokered computing

Non-deterministic event brokered computing
复制标题

非确定性事件代理计算

DOI:
10.1145/3535044.3535055
复制
发表时间:
2022
期刊:
--
影响因子:
--
通讯作者:
Brown A
Brown A
中科院分区:
--
文献类型:
--
作者:
Brown A

文献摘要

参考文献

相似文献

本文回顾了大规模微并行计算系统POETS(偏序事件触发系统),并说明了它在加速高要求应用方面的潜力。受益于POETS的应用领域包括可以离散为网格的物理系统的模拟。问题图分布在一个大型计算网格上;每个网格顶点都包含一个处理器--一个基于FPGA的RISC-V线程,支持我们原型中的定制指令--以及少量的本地问题状态数据。没有任何类型的中央监督者,处理器除了自己的内存之外无法看到内存。问题图顶点与邻居交互,通过发送异步分组来发送状态更改。数据包大小固定--目前为字节--硬件通信基础设施非常快。应用程序可以使用异步“数据包风暴”方法;使用硬件空闲屏障同步运行,或者以全局异步、本地同步的方式运行。结果表明,与传统系统相比,挂钟加速比和功耗都有显著提升:对于一个应用程序,我们显示的速度比基于传统CPU的系统快40倍;与多GPU系统相比,POETS集群的速度快26%,能效高60%,能效高34%。
This paper reviews the massively micro-parallel compute system POETS (Partially Ordered Event Triggered System) and illustrates its potential for speeding up demanding applications. Application domains that benefit from POETS include simulations of physical systems that can be discretised as a mesh. The problem graph is distributed over a large compute mesh; each mesh vertex contains a processor – an FPGA-based RISC-V thread supporting custom instructions in our prototype – and a small amount of local problem state data. There is no central overseer of any sort and processors cannot see memory besides their own. A problem graph vertex interacts with a neighbour to send a state change by sending an asynchronous packet. The packets are fixed size and small – currently 64 bytes – and the hardware communications infrastructure is very fast. Applications can use an asynchronous ‘packet storm’ approach; run synchronously using a hardware idle barrier or run in a globally asynchronous, locally synchronous manner. Results show significant wallclock speedup and power consumption improvement over conventional systems: for one application we show a 40-fold speedup over a conventional CPU-based system; versus a multi-GPU system, the POETS cluster is 26% faster, 60% more power efficient, and 34% more energy efficient.
网络处理器模板的自定义增强功能
DOI: --
发表时间: 2021
期刊: IEEE Computer Society Annual Symposium on VLSI
影响因子: --
作者:
T. Todman;W. Luk
通讯作者: W. Luk
POETS:尖峰神经网络的并行集群架构
DOI: 10.18178/ijmlc.2021.11.4.1048
发表时间: 2021
期刊: International Journal of Machine Learning and Computing
影响因子: --
作者:
Mahyar Shahsavari;Jonathan Beaumont;David B. Thomas;A. Brown
通讯作者: A. Brown
在基于事件的可重构集群架构上使用基于奖励的 STDP 学习进行神经形态设计
DOI: 10.1145/3477145.3477151
发表时间: 2021
期刊: International Conference on Neuromorphic Systems 2021
影响因子: --
作者:
Mahyar Shahsavari;David B. Thomas;Andrew D. Brown;W. Luk
通讯作者: W. Luk
DOI: --
发表时间: 2022
期刊: Trans. Petri Nets Other Model. Concurr.
影响因子: --
作者:
A. Rafiev;Jordan Morris;F. Xia;A. Yakovlev;Matthew Naylor;S. Moore;David B. Thomas;Graeme M. Bragg;Mark L. Vousden;Andrew D. Brown
通讯作者: Andrew D. Brown
Tinsel:FPGA 集群的多线程叠加
DOI: --
发表时间: 2019
期刊: International Conference on Field-Programmable Logic and Applications
影响因子: --
作者:
Matthew Naylor;S. Moore;David B. Thomas
通讯作者: David B. Thomas