DOCSLIB.ORG

Designing Low Power and High Performance Network-On-Chip Communication Architectures for Nanometer Socs

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Designing Low Power and High Performance Network-On-Chip Communication Architectures for Nanometer Socs Designing Low Power and High Performance Network-on-Chip Communication Architectures for Nanometer SoCs DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University By Gursharan Reehal, B.S., M.S. Graduate Program in Electrical and Computer Engineering The Ohio State University 2012 Dissertation Committee: Prof. Mohammed Ismail El-Naggar, Advisor Prof. Steve Bibyk Prof. Joanne DeGroat c Copyright by Gursharan Reehal 2012 ABSTRACT Network-on-Chip (NoC) communication architectures have been recognized as the most scalable and efficient solution for on chip communication challenges in the multi-core era. Diverse demanding applications coupled with the ability to integrate billions of transistors on a single chip are some of the main driving forces behind ever increasing performance requirements towards the level that requires several tens to over a hundred of cores per chip, with aggregate performance exceeding one trillion operations per second. Such tera-scale many-core processors will be highly integrated System-on-Chip designs (SoC) containing a variety of on-chip storage elements, mem- ory controllers, and input/output (I/O) functional blocks. Small scale multicore pro- cessors so far have been a great commercial success and found applicability in high bandwidth, computer intensive applications including high performance, throughput oriented, scientific computing, high performance graphics and 3-D immersive visual interfaces, as well as in decision and support systems. Systems using multi-core pro- cessors are now the norm rather than the exception. As the number of cores or components integrated into a single system is keep increasing, the design of on-chip communication architecture is becoming more chal- lenging. The increasing number of components in a system translates into more ii inter-component communication that must be handled by the on-chip communica- tion infrastructure. It's not surprising to see that leading-edge design teams search- ing for better solutions as multi-core SoCs continue to evolve. Future system-on-chip (SoC) designs require predictable, scalable and reusable on-chip communication ar- chitectures to increase reliability and productivity. Current bus-based interconnect architectures are inherently non-scalable, less adaptable for reuse and their reliability decreases with system size. NoC communication guarantees scalability, high-speed, high-bandwidth commu- nication with minimal wiring overhead and routing issues. NoCs are layered, packet- based on-chip communication networks integrated onto a single chip and their opera- tion is based on the operating principle of macro networks. NoC consists of resources and switches that are directly connected in a way that resources are able to com- municate with each other by sending messages. The proficiency of a NoC to meet its design goals and budget requirements for the target application depends on its design. Often, these design goals conflict and trade-off with each other. The multi- dimensional pull of design constraints in addition to technology scaling complicates the process of NoC design in many aspects, as they are expected to support high performance and reliability along with low cost, smaller area, less time-to-market and lower power consumption. To aid the process, this research presents design method- ologies to achieve low power and high performance NoC communication architectures for nanometer SoCs. In NoCs, interconnects play a crucial role in the overall system performance and can have a large impact on the total power consumption, wiring area and achievable system performance. The effect of technology scaling on the NoC interconnects is iii studied and an improved design flow is presented. The influence of technology node, die size, number of components on the power consumption by the NoC interconnects is analyzed. The success of NoC heavily depends on its power budget. As CMOS technology continues to scale, power aware design is more important than ever before, especially for the designs targeted towards low power applications, however in large scale NoCs the power consumption can increase beyond acceptable limits. Designing low power NoC is therefore extremely important especially for larger SoCs designs. The elevation of power to a first-class design constraint requires that power estimations are done at the same time as the performance studies in the design flow. In NoC, one method to have a power aware design is to consider the impact of the architectural choices on the power in the early stages of design process. In this research an efficient design methodology based on the layout and power models is presented to have rough power estimates in the early stages of design cycle. The impact of die size and number of IPs on the power consumed by different NoC architectures is evaluated. Additionally, as multi-core SoCs continue to evolve, Globally Asynchronous Lo- cally Synchronous (GALS) design techniques have been suggested as a potential solu- tion in larger and faster SoC designs to avoid the problems related to synchronization and clock skew. These multi-core SoCs will operate using GALS paradigm, where each core can operate in a separate clock domain. In this research a study on the power efficiency between Synchronous and Asynchronous NoC architecures is presented. Asynchronous NoC architecture is shown to consume less power, when activity factor of data transfers between two switches is within certain range. Asynchronous designs are more power efficient, as the need for clock distribution is eliminated. iv Dedicated to my mother, for her love, support and encouragement. v ACKNOWLEDGMENTS I would like to thank many remarkable people who have supported and encouraged me during my time at The Ohio State University. First and foremost, I would like to thank my advisor, Prof. Mohammed Ismail for his continued guidance, support and sustained encouragement throughout my graduate study. He always encouraged me to think independently and to define my own research agenda, allowing me to develop the skills necessary to do research. His significant effort in bringing the Synopsys Software, an industry standard EDA tool for research in the area of Digital VLSI and henceforth enriching the quality of education and student research experience here at The Ohio State University is highly remarkable. Without his commitment and encouragement, this dissertation would not have been possible. The experience with Prof. Ismail will always be highly regarded. I would like to thank Prof. Steve Bibyk for his early guidance, mentorship, en- couragement and support to pursue the PhD program. I am thankful for his valuable discussions throughout my time in the ECE department and for being my MS advi- sor. It truly has been a great experience working with him. I am also very grateful to Prof. Joanne DeGroat for her support, guidance and for kindly serving on my PhD exam committees. I am thankful to her for treating me as a member of her own group. vi I am thankful for the opportunity to do some research with Mohamed Abd Ghany, who is affiliated with the German University Cairo in Cairo Egypt. He has been a very helpful friend and a great source of guidance in the work on Asynchronous NoCs. I am grateful for his insight in the asynchronous digital design and for his support. I enjoyed working with him. I am honored to call myself a member of the VLSI Lab. I want to thank my fellow colleagues Amneh Akour, Sleiman Bou Sleiman, John Hu, Sidharth Balasubramanian, Yiqiao Lin, Feiran Lei and Samantha Yoder for their friendship, discussions and valuable guidance. In particular, I would like to thank Amneh Akour, who has been a great source of sage advice in the times, when I needed it the most. It truly has been a privilege for me to work with them in the VLSI lab and I always enjoyed and cherished their company. I would also like to thank many people in the ECE department. Stephanie Muldrow, Carol Duhigg, Tricia Toothman, Vincent Juodvalkis, Aaron Aufderheide, Don Gibb, and Edwin Lim, who work very hard and diligently behind the scenes, to make this department a wonderful place for graduate students. Their friendly and helping nature ease the stress of a graduate student and who are always willing to help and guide students in our department. Life as a graduate student would not be possible without the help of family and friends. My deepest regards goes to my mother, to whom I dedicate this work. She always encouraged me and supported me to pursue higher education. She worked very hard in making sure, education is a priority. She always emphasized the importance of education, and encouraged me to pursue the PhD program. She has been a great source of inspiration in my life and education. I am deeply thankful to her for her vii endless love, care, support, wisdom...and basically everything. Your love and faith in me has made all the difference...love you Mom! My final thanks goes to the reason I am here and able to do this work, my creator. I am thankful for the talent and opportunities I have been given and the strength to undertake this task and see it to completion! viii VITA 1996 . .B.S. Electrical Engineering 1998 . .M.S. Electrical Engineering 2007- 2009 . .Graduate Teaching Associate, The Ohio State University. 2010 . .Graduate Technical Intern, Intel Corporation 2010 . .Graduate Technical Research Intern, Intel Labs FIELDS OF STUDY Major Field: Electrical and Computer Engineering ix
Load more

Recommended publications
  • A Survey of Network Performance Monitoring Tools
    A Survey of Network Performance Monitoring Tools Travis Keshav -- [email protected] Abstract In today's world of networks, it is not enough simply to have a network; assuring its optimal performance is key. This paper analyzes several facets of Network Performance Monitoring, evaluating several motivations as well as examining many commercial and public domain products. Keywords: network performance monitoring, application monitoring, flow monitoring, packet capture, sniffing, wireless networks, path analysis, bandwidth analysis, network monitoring platforms, Ethereal, Netflow, tcpdump, Wireshark, Ciscoworks Table Of Contents 1. Introduction 2. Application & Host-Based Monitoring 2.1 Basis of Application & Host-Based Monitoring 2.2 Public Domain Application & Host-Based Monitoring Tools 2.3 Commercial Application & Host-Based Monitoring Tools 3. Flow Monitoring 3.1 Basis of Flow Monitoring 3.2 Public Domain Flow Monitoring Tools 3.3 Commercial Flow Monitoring Protocols 4. Packet Capture/Sniffing 4.1 Basis of Packet Capture/Sniffing 4.2 Public Domain Packet Capture/Sniffing Tools 4.3 Commercial Packet Capture/Sniffing Tools 5. Path/Bandwidth Analysis 5.1 Basis of Path/Bandwidth Analysis 5.2 Public Domain Path/Bandwidth Analysis Tools 6. Wireless Network Monitoring 6.1 Basis of Wireless Network Monitoring 6.2 Public Domain Wireless Network Monitoring Tools 6.3 Commercial Wireless Network Monitoring Tools 7. Network Monitoring Platforms 7.1 Basis of Network Monitoring Platforms 7.2 Commercial Network Monitoring Platforms 8. Conclusion 9. References and Acronyms 1.0 Introduction http://www.cse.wustl.edu/~jain/cse567-06/ftp/net_perf_monitors/index.html 1 of 20 In today's world of networks, it is not enough simply to have a network; assuring its optimal performance is key.
    [Show full text]
  • Development of a Dynamically Extensible Spinnaker Chip Computing Module
    FACULDADE DE ENGENHARIA DA UNIVERSIDADE DO PORTO Development of a Dynamically Extensible SpiNNaker Chip Computing Module Rui Emanuel Gonçalves Calado Araújo Master in Electrical and Computers Engineering Supervisor: Jörg Conradt Co-Supervisor: Diamantino Freitas January 27, 2014 Resumo O projeto SpiNNaker desenvolveu uma arquitetura que é capaz de criar um sistema com mais de um milhão de núcleos, com o objetivo de simular mais de um bilhão de neurónios em tempo real biológico. O núcleo deste sistema é o "chip" SpiNNaker, um multiprocessador System-on-Chip com um elevado nível de interligação entre as suas unidades de processamento. Apesar de ser uma plataforma de computação com muito potencial, até para aplicações genéricas, atualmente é ape- nas disponibilizada em configurações fixas e requer uma estação de trabalho, como uma máquina tipo "desktop" ou "laptop" conectada através de uma conexão Ethernet, para a sua inicialização e receber o programa e os dados a processar. No sentido de tirar proveito das capacidades do "chip" SpiNNaker noutras áreas, como por exemplo, na área da robótica, nomeadamente no caso de robots voadores ou de tamanho pequeno, uma nova solução de hardware com software configurável tem de ser projetada de forma a poder selecionar granularmente a quantidade do poder de processamento. Estas novas capacidades per- mitem que a arquitetura SpiNNaker possa ser utilizada em mais aplicações para além daquelas para que foi originalmente projetada. Esta dissertação apresenta um módulo de computação dinamicamente extensível baseado em "chips" SpiNNaker com a finalidade de ultrapassar as limitações supracitadas das máquinas SpiN- Naker atualmente disponíveis. Esta solução consiste numa única placa com um microcontrolador, que emula um "chip" SpiNNaker com uma ligação Ethernet, acessível através de uma porta série e com um "chip" SpiNNaker.
    [Show full text]
  • An Introduction
    This chapter is from Social Media Mining: An Introduction. By Reza Zafarani, Mohammad Ali Abbasi, and Huan Liu. Cambridge University Press, 2014. Draft version: April 20, 2014. Complete Draft and Slides Available at: http://dmml.asu.edu/smm Chapter 10 Behavior Analytics What motivates individuals to join an online group? When individuals abandon social media sites, where do they migrate to? Can we predict box office revenues for movies from tweets posted by individuals? These questions are a few of many whose answers require us to analyze or predict behaviors on social media. Individuals exhibit different behaviors in social media: as individuals or as part of a broader collective behavior. When discussing individual be- havior, our focus is on one individual. Collective behavior emerges when a population of individuals behave in a similar way with or without coordi- nation or planning. In this chapter we provide examples of individual and collective be- haviors and elaborate techniques used to analyze, model, and predict these behaviors. 10.1 Individual Behavior We read online news; comment on posts, blogs, and videos; write reviews for products; post; like; share; tweet; rate; recommend; listen to music; and watch videos, among many other daily behaviors that we exhibit on social media. What are the types of individual behavior that leave a trace on social media? We can generally categorize individual online behavior into three cate- gories (shown in Figure 10.1): 319 Figure 10.1: Individual Behavior. 1. User-User Behavior. This is the behavior individuals exhibit with re- spect to other individuals. For instance, when befriending someone, sending a message to another individual, playing games, following, inviting, blocking, subscribing, or chatting, we are demonstrating a user-user behavior.
    [Show full text]
  • An Architecture and Compiler for Scalable On-Chip Communication
    IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. XX, NO. Y, MONTH 2004 1 An Architecture and Compiler for Scalable On-Chip Communication Jian Liang, Student Member, IEEE, Andrew Laffely, Sriram Srinivasan, and Russell Tessier, Member, IEEE Abstract— tion of communication resources. Significant amounts of arbi- A dramatic increase in single chip capacity has led to a tration across even a small number of components can quickly revolution in on-chip integration. Design reuse and ease-of- form a performance bottleneck, especially for data-intensive, implementation have became important aspects of the design pro- cess. This paper describes a new scalable single-chip communi- stream-based computation. This issue is made more complex cation architecture for heterogeneous resources, adaptive System- by the need to compile high-level representations of applica- On-a-Chip (aSOC), and supporting software for application map- tions to SoC environments. The heterogeneous nature of cores ping. This architecture exhibits hardware simplicity and opti- in terms of clock speed, resources, and processing capability mized support for compile-time scheduled communication. To il- makes cost modeling difficult. Additionally, communication lustrate the benefits of the architecture, four high-bandwidth sig- nal processing applications including an MPEG-2 video encoder modeling for interconnection with long wires and variable arbi- and a Doppler radar processor have been mapped to a prototype tration protocols limits performance predictability required by aSOC device using our design mapping technology. Through ex- computation scheduling. perimentation it is shown that aSOC communication outperforms Our platform for on-chip interconnect, adaptive System-On- a hierarchical bus-based system-on-chip (SoC) approach by up to a-Chip (aSOC), is a modular communications architecture.
    [Show full text]
  • Three-Dimensional Integrated Circuit Design: EDA, Design And
    Integrated Circuits and Systems Series Editor Anantha Chandrakasan, Massachusetts Institute of Technology Cambridge, Massachusetts For other titles published in this series, go to http://www.springer.com/series/7236 Yuan Xie · Jason Cong · Sachin Sapatnekar Editors Three-Dimensional Integrated Circuit Design EDA, Design and Microarchitectures 123 Editors Yuan Xie Jason Cong Department of Computer Science and Department of Computer Science Engineering University of California, Los Angeles Pennsylvania State University [email protected] [email protected] Sachin Sapatnekar Department of Electrical and Computer Engineering University of Minnesota [email protected] ISBN 978-1-4419-0783-7 e-ISBN 978-1-4419-0784-4 DOI 10.1007/978-1-4419-0784-4 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2009939282 © Springer Science+Business Media, LLC 2010 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com) Foreword We live in a time of great change.
    [Show full text]
  • On-Chip Interconnect Schemes for Reconfigurable System-On-Chip
    On-chip Interconnect Schemes for Reconfigurable System-on-Chip Andy S. Lee, Neil W. Bergmann. School of ITEE, The University of Queensland, Brisbane Australia {andy, n.bergmann} @itee.uq.edu.au ABSTRACT On-chip communication architectures can have a great influence on the speed and area of System-on-Chip designs, and this influence is expected to be even more pronounced on reconfigurable System-on-Chip (rSoC) designs. To date, little research has been conducted on the performance implications of different on-chip communication architectures for rSoC designs. This paper motivates the need for such research and analyses current and proposed interconnect technologies for rSoC design. The paper also describes work in progress on implementation of a simple serial bus and a packet-switched network, as well as a methodology for quantitatively evaluating the performance of these interconnection structures in comparison to conventional buses. Keywords: FPGAs, Reconfigurable Logic, System-on-Chip 1. INTRODUCTION System-on-chip (SoC) technology has evolved as the predominant circuit design methodology for custom ASICs. SoC technology moves design from the circuit level to the system level, concentrating on the selection of appropriate pre-designed IP Blocks, and their interconnection into a complete system. However, modern ASIC design and fabrication are expensive. Design tools may cost many hundreds of thousands of dollars, while tooling and mask costs for large SoC designs now approach $1million. For low volume applications, and especially for research and development projects in universities, reconfigurable System-on-Chip (rSoC) technology is more cost effective. Like conventional SoC design, rSoC involves the assembly of predefined IP blocks (such as processors and peripherals) and their interconnection.
    [Show full text]
  • Learning to Predict End-To-End Network Performance
    CORE Metadata, citation and similar papers at core.ac.uk Provided by BICTEL/e - ULg UNIVERSITÉ DE LIÈGE Faculté des Sciences Appliquées Institut d’Électricité Montefiore RUN - Research Unit in Networking Learning to Predict End-to-End Network Performance Yongjun Liao Thèse présentée en vue de l’obtention du titre de Docteur en Sciences de l’Ingénieur Année Académique 2012-2013 ii iii Abstract The knowledge of end-to-end network performance is essential to many Internet applications and systems including traffic engineering, content distribution networks, overlay routing, application- level multicast, and peer-to-peer applications. On the one hand, such knowledge allows service providers to adjust their services according to the dynamic network conditions. On the other hand, as many systems are flexible in choosing their communication paths and targets, knowing network performance enables to optimize services by e.g. intelligent path selection. In the networking field, end-to-end network performance refers to some property of a network path measured by various metrics such as round-trip time (RTT), available bandwidth (ABW) and packet loss rate (PLR). While much progress has been made in network measurement, a main challenge in the acquisition of network performance on large-scale networks is the quadratical growth of the measurement overheads with respect to the number of network nodes, which ren- ders the active probing of all paths infeasible. Thus, a natural idea is to measure a small set of paths and then predict the others where there are no direct measurements. This understanding has motivated numerous research on approaches to network performance prediction.
    [Show full text]
  • NVIDIA CUDA on IBM POWER8: Technical Overview, Software Installation, and Application Development
    Redpaper Dino Quintero Wei Li Wainer dos Santos Moschetta Mauricio Faria de Oliveira Alexander Pozdneev NVIDIA CUDA on IBM POWER8: Technical overview, software installation, and application development Overview The exploitation of general-purpose computing on graphics processing units (GPUs) and modern multi-core processors in a single heterogeneous parallel system has proven highly efficient for running several technical computing workloads. This applied to a wide range of areas such as chemistry, bioinformatics, molecular biology, engineering, and big data analytics. Recently launched, the IBM® Power System S824L comes into play to explore the use of the NVIDIA Tesla K40 GPU, combined with the latest IBM POWER8™ processor, providing a unique technology platform for high performance computing. This IBM Redpaper™ publication discusses the installation of the system, and the development of C/C++ and Java applications using the NVIDIA CUDA platform for IBM POWER8. Note: CUDA stands for Compute Unified Device Architecture. It is a parallel computing platform and programming model created by NVIDIA and implemented by the GPUs that they produce. The following topics are covered: Advantages of NVIDIA on POWER8 The IBM Power Systems S824L server Software stack System monitoring Application development Tuning and debugging Application examples © Copyright IBM Corp. 2015. All rights reserved. ibm.com/redbooks 1 Advantages of NVIDIA on POWER8 The IBM and NVIDIA partnership was announced in November 2013, for the purpose of integrating IBM POWER®-based systems with NVIDIA GPUs, and enablement of GPU-accelerated applications and workloads. The goal is to deliver higher performance and better energy efficiency to companies and data centers. This collaboration produced its initial results in 2014 with: The announcement of the first IBM POWER8 system featuring NVIDIA Tesla GPUs (IBM Power Systems™ S824L).
    [Show full text]
  • Designing Vertical Bandwidth Reconfigurable 3D Nocs for Many
    Designing Vertical Bandwidth Reconfigurable 3D NoCs for Many Core Systems Qiaosha Zou∗, Jia Zhany, Fen Gez, Matt Poremba∗, Yuan Xiey ∗ Computer Science and Engineering, The Pennsylvania State University, USA y Electrical and Computer Engineering, University of California, Santa Barbara, USA z Nanjing University of Aeronautics and Astronautics, China Email: ∗fqszou, [email protected], yfjzhan, [email protected], z [email protected] Abstract—As the number of processing elements increases in a single Routers chip, the interconnect backbone becomes more and more stressed when Core 1 Cache/ Cache/Memory serving frequent memory and cache accesses. Network-on-Chip (NoC) Memory has emerged as a potential solution to provide a flexible and scalable Core 0 interconnect in a planar platform. In the mean time, three-dimensional TSVs (3D) integration technology pushes circuit design beyond Moore’s law and provides short vertical connections between different layers. As a result, the innovative solution that combines 3D integrations and NoC designs can further enhance the system performance. However, due Core 0 Core 1 to the unpredictable workload characteristics, NoC may suffer from intermittent congestions and channel overflows, especially when the (a) (b) network bandwidth is limited by the area and energy budget. In this work, we explore the performance bottlenecks in 3D NoC, and then leverage Fig. 1. Overview of core-to-cache/memory 3D stacking. (a). Cores are redundant TSVs, which are conventionally used for fault tolerance only, allocated in all layers with part of cache/memory; (b). All cores are located as vertical links to provide additional channel bandwidth for instant in the same layers while cache/memory in others.
    [Show full text]
  • A Social Network Matrix for Implicit and Explicit Social Network Plates
    A Social Network Matrix for Implicit and Explicit Social Network Plates Wei Zhou1;4, Wenjing Duan2, Selwyn Piramuthu3;4 1Information & Operations Management, ESCP Europe, Paris, France 2Information Systems and Technology Management, George Washington University, U.S.A. 3Information Systems and Operations Management, University of Florida, U.S.A. Gainesville, Florida 32611-7169, USA 4RFID European Lab, Paris, France. [email protected], [email protected], selwyn@ufl.edu Abstract A majority of social network research deals with explicitly formed social networks. Although only rarely acknowledged for its existence, we believe that implicit social networks play a significant role in the overall dynamics of social networks. We propose a framework to evalu- ate the dynamics and characteristics of a set of explicit and associated implicit social networks. Specifically, we propose a social network ma- trix to measure the implicit relationships among the entities in various social networks. We also derive several indicators to characterize the dynamics in online social networks. We proceed by incorporating im- plicit social networks in a traditional network flow context to evaluate key network performance indicators such as the lowest communication cost, maximum information flow, and the budgetary constraints. Keywords: Implicit Social Network, Online Social Network 1 1 Introduction In recent years, several online social network platforms have witnessed huge public attention from social and financial perspectives. However, there are different facets to this interest. Facebook, for example, has gained a large set of users but failed to excel on profitability. Online social networks can be broadly classified as explicit and implicit social networks. Explicit social networks (e.g., Facebook, LinkedIn, Twitter, and MySpace) are where the users define the network by explicitly connect- ing with other users, possibly, but not necessarily, based on shared interests.
    [Show full text]
  • Openpower AI CERN V1.Pdf
    Moore’s Law Processor Technology Firmware / OS Linux Accelerator sSoftware OpenStack Storage Network ... Price/Performance POWER8 2000 2020 DRAM Memory Chips Buffer Power8: Up to 12 Cores, up to 96 Threads L1, L2, L3 + L4 Caches Up to 1 TB per socket https://www.ibm.com/blogs/syst Up to 230 GB/s sustained memory ems/power-systems- openpower-enable- bandwidth acceleration/ System System Memory Memory 115 GB/s 115 GB/s POWER8 POWER8 CPU CPU NVLink NVLink 80 GB/s 80 GB/s P100 P100 P100 P100 GPU GPU GPU GPU GPU GPU GPU GPU Memory Memory Memory Memory GPU PCIe CPU 16 GB/s System bottleneck Graphics System Memory Memory IBM aDVantage: data communication and GPU performance POWER8 + 78 ms Tesla P100+NVLink x86 baseD 170 ms GPU system ImageNet / Alexnet: Minibatch size = 128 ADD: Coherent Accelerator Processor Interface (CAPI) FPGA CAPP PCIe POWER8 Processor ...FPGAs, networking, memory... Typical I/O MoDel Flow Copy or Pin MMIO Notify Poll / Int Copy or Unpin Ret. From DD DD Call Acceleration Source Data Accelerator Completion Result Data Completion Flow with a Coherent MoDel ShareD Mem. ShareD Memory Acceleration Notify Accelerator Completion Focus on Enterprise Scale-Up Focus on Scale-Out and Enterprise Future Technology and Performance DriVen Cost and Acceleration DriVen Partner Chip POWER6 Architecture POWER7 Architecture POWER8 Architecture POWER9 Architecture POWER10 POWER8/9 2007 2008 2010 2012 2014 2016 2017 TBD 2018 - 20 2020+ POWER6 POWER6+ POWER7 POWER7+ POWER8 POWER8 P9 SO P9 SU P9 SO 2 cores 2 cores 8 cores 8 cores 12 cores w/ NVLink
    [Show full text]
  • AXI Reference Guide
    AXI Reference Guide [Guide Subtitle] [optional] UG761 (v13.4) January 18, 2012 [optional] Xilinx is providing this product documentation, hereinafter “Information,” to you “AS IS” with no warranty of any kind, express or implied. Xilinx makes no representation that the Information, or any particular implementation thereof, is free from any claims of infringement. You are responsible for obtaining any rights you may require for any implementation based on the Information. All specifications are subject to change without notice. XILINX EXPRESSLY DISCLAIMS ANY WARRANTY WHATSOEVER WITH RESPECT TO THE ADEQUACY OF THE INFORMATION OR ANY IMPLEMENTATION BASED THEREON, INCLUDING BUT NOT LIMITED TO ANY WARRANTIES OR REPRESENTATIONS THAT THIS IMPLEMENTATION IS FREE FROM CLAIMS OF INFRINGEMENT AND ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Except as stated herein, none of the Information may be copied, reproduced, distributed, republished, downloaded, displayed, posted, or transmitted in any form or by any means including, but not limited to, electronic, mechanical, photocopying, recording, or otherwise, without the prior written consent of Xilinx. © Copyright 2012 Xilinx, Inc. XILINX, the Xilinx logo, Virtex, Spartan, Kintex, Artix, ISE, Zynq, and other designated brands included herein are trademarks of Xilinx in the United States and other countries. All other trademarks are the property of their respective owners. ARM® and AMBA® are registered trademarks of ARM in the EU and other countries. All other trademarks are the property of their respective owners. Revision History The following table shows the revision history for this document: . Date Version Description of Revisions 03/01/2011 13.1 Second Xilinx release.
    [Show full text]