Sirius

Author:

Hauswald Johann1,Laurenzano Michael A.1,Zhang Yunqi1,Li Cheng1,Rovinski Austin1,Khurana Arjun1,Dreslinski Ronald G.1,Mudge Trevor1,Petrucci Vinicius2,Tang Lingjia1,Mars Jason1

Affiliation:

1. University of Michigan, Ann Arbor, MI, USA

2. Federal University of Bahia, Bahia, Brazil

Abstract

As user demand scales for intelligent personal assistants (IPAs) such as Apple's Siri, Google's Google Now, and Microsoft's Cortana, we are approaching the computational limits of current datacenter architectures. It is an open question how future server architectures should evolve to enable this emerging class of applications, and the lack of an open-source IPA workload is an obstacle in addressing this question. In this paper, we present the design of Sirius, an open end-to-end IPA web-service application that accepts queries in the form of voice and images, and responds with natural language. We then use this workload to investigate the implications of four points in the design space of future accelerator-based server architectures spanning traditional CPUs, GPUs, manycore throughput co-processors, and FPGAs. To investigate future server designs for Sirius, we decompose Sirius into a suite of 7 benchmarks (Sirius Suite) comprising the computationally intensive bottlenecks of Sirius. We port Sirius Suite to a spectrum of accelerator platforms and use the performance and power trade-offs across these platforms to perform a total cost of ownership (TCO) analysis of various server design points. In our study, we find that accelerators are critical for the future scalability of IPA services. Our results show that GPU- and FPGA-accelerated servers improve the query latency on average by 10x and 16x. For a given throughput, GPU- and FPGA-accelerated servers can reduce the TCO of datacenters by 2.6x and 1.4x, respectively.

Publisher

Association for Computing Machinery (ACM)

Reference71 articles.

1. Apple's Siri. https://www.apple.com/ios/siri/. Apple's Siri. https://www.apple.com/ios/siri/.

2. Google's Google Now. http://www.google.com/landing/now/. Google's Google Now. http://www.google.com/landing/now/.

3. Microsoft's Cortana. http://www.windowsphone.com/en-us/features-8--1. Microsoft's Cortana. http://www.windowsphone.com/en-us/features-8--1.

4. Smartphone OS Market Share Q1 2014. http://www.idc.com/prodserv/smartphone-os-market-share.jsp. Smartphone OS Market Share Q1 2014. http://www.idc.com/prodserv/smartphone-os-market-share.jsp.

5. Google's Android Wear. www.android.com/wear/. Google's Android Wear. www.android.com/wear/.

Cited by 6 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. A scalable and adaptable allocation framework for heterogeneous resources in a large cluster environment;Concurrency and Computation: Practice and Experience;2019-11-20

2. Framework for Creating Audio Games for Intelligent Personal Assistants;Advances in Human Factors in Wearable Technologies and Game Design;2017-06-14

3. Building a K-Pop knowledge graph using an entertainment ontology;Knowledge Management Research & Practice;2017-05

4. Generating Configurable Hardware from Parallel Patterns;ACM SIGOPS Operating Systems Review;2016-03-25

5. Towards Safe and Secure Autonomous and Cooperative Vehicle Ecosystems;Proceedings of the 2nd ACM Workshop on Cyber-Physical Systems Security and Privacy - CPS-SPC '16;2016

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3