Interconnection Generation for System-on-Chip Design and Design Space Exploration



Similar documents
Applying the Benefits of Network on a Chip Architecture to FPGA System Design

Lizy Kurian John Electrical and Computer Engineering Department, The University of Texas as Austin

Qsys and IP Core Integration

Design and Implementation of an On-Chip timing based Permutation Network for Multiprocessor system on Chip

Distributed Elastic Switch Architecture for efficient Networks-on-FPGAs

Architectural Level Power Consumption of Network on Chip. Presenter: YUAN Zheng

From Bus and Crossbar to Network-On-Chip. Arteris S.A.

SoC IP Interfaces and Infrastructure A Hybrid Approach

Computer Systems Structure Input/Output

Optimizing Configuration and Application Mapping for MPSoC Architectures

Learning Outcomes. Simple CPU Operation and Buses. Composition of a CPU. A simple CPU design

Interconnection Networks

Switched Interconnect for System-on-a-Chip Designs

DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA EFFICIENT ROUTER DESIGN FOR NETWORK ON CHIP

Design and Implementation of an On-Chip Permutation Network for Multiprocessor System-On-Chip

LogiCORE IP AXI Performance Monitor v2.00.a

Design and Verification of Nine port Network Router

Hardware Implementation of Improved Adaptive NoC Router with Flit Flow History based Load Balancing Selection Strategy

Switch Fabric Implementation Using Shared Memory

AXI Performance Monitor v5.0

Maximizing Server Storage Performance with PCI Express and Serial Attached SCSI. Article for InfoStor November 2003 Paul Griffith Adaptec, Inc.

Chapter 13 Selected Storage Systems and Interface

AMD Opteron Quad-Core

Arbitration and Switching Between Bus Masters

What is a System on a Chip?

Computer Architecture

DDR subsystem: Enhancing System Reliability and Yield

How To Design A Single Chip System Bus (Amba) For A Single Threaded Microprocessor (Mma) (I386) (Mmb) (Microprocessor) (Ai) (Bower) (Dmi) (Dual

CH.1. Lecture # 2. Computer Networks and the Internet. Eng. Wafaa Audah. Islamic University of Gaza. Faculty of Engineering

SOCWIRE: A SPACEWIRE INSPIRED FAULT TOLERANT NETWORK-ON-CHIP FOR RECONFIGURABLE SYSTEM-ON-CHIP DESIGNS

A Generic Network Interface Architecture for a Networked Processor Array (NePA)

What is a bus? A Bus is: Advantages of Buses. Disadvantage of Buses. Master versus Slave. The General Organization of a Bus

Testing of Digital System-on- Chip (SoC)

Asynchronous Bypass Channels

Power Reduction Techniques in the SoC Clock Network. Clock Power

Asynchronous IC Interconnect Network Design and Implementation Using a Standard ASIC Flow

Using a Generic Plug and Play Performance Monitor for SoC Verification

Voice over Internet Protocol (VoIP) systems can be built up in numerous forms and these systems include mobile units, conferencing units and

SOC architecture and design

PCI Express: The Evolution to 8.0 GT/s. Navraj Nandra, Director of Marketing Mixed-Signal and Analog IP, Synopsys

Overview of Network Hardware and Software. CS158a Chris Pollett Jan 29, 2007.

Operating System Support for Multiprocessor Systems-on-Chip

Managing High-Speed Clocks

Interconnection Networks Programmierung Paralleler und Verteilter Systeme (PPV)

10/100/1000Mbps Ethernet MAC with Protocol Acceleration MAC-NET Core with Avalon Interface

Outline. Introduction. Multiprocessor Systems on Chip. A MPSoC Example: Nexperia DVP. A New Paradigm: Network on Chip

Design of a High Speed Communications Link Using Field Programmable Gate Arrays

Atmel Norway XMEGA Introduction

On-Chip Communications Network Report

Intel PCI and PCI Express*

Lecture 18: Interconnection Networks. CMU : Parallel Computer Architecture and Programming (Spring 2012)

8-ch RAID0 Design by using SATA Host IP Manual Rev1.0 9-Jun-15

Reconfigurable Architecture Requirements for Co-Designed Virtual Machines

The proliferation of the raw processing

DESIGN AND VERIFICATION OF LSR OF THE MPLS NETWORK USING VHDL

Kingdom of Saudi Arabia King Saud University

AND8336. Design Examples of On Board Dual Supply Voltage Logic Translators. Prepared by: Jim Lepkowski ON Semiconductor.

OpenSoC Fabric: On-Chip Network Generator

NETWORKING TECHNOLOGIES

DDR3 DIMM Slot Interposer

Memory ICS 233. Computer Architecture and Assembly Language Prof. Muhamed Mudawar

On-Chip Interconnection Networks Low-Power Interconnect

Cray Gemini Interconnect. Technical University of Munich Parallel Programming Class of SS14 Denys Sobchyshak

IT4405 Computer Networks (Compulsory)

ZigBee Technology Overview

SpW-10X Network Performance Testing. Peter Mendham, Jon Bowyer, Stuart Mills, Steve Parkes. Space Technology Centre University of Dundee

Local Area Networks transmission system private speedy and secure kilometres shared transmission medium hardware & software

CSE3214 Computer Network Protocols and Applications. Chapter 1 Examples and Homework Problems

VME Data Acquisition System: Fundamentals and Beyond. Abhinav Kumar Bhabha Atomic Research Centre, Mumbai March 2011

Input / Ouput devices. I/O Chapter 8. Goals & Constraints. Measures of Performance. Anatomy of a Disk Drive. Introduction - 8.1

Designing Reliable IP/MPLS Core Transport Networks

FlexPath Network Processor

Chapter 7: Computer Networks, the Internet, and the World Wide Web. Invitation to Computer Science, C++ Version, Third Edition

Communicating with devices

Communication Networks. MAP-TELE 2011/12 José Ruela

The implementation and performance/cost/power analysis of the network security accelerator on SoC applications

Achieving Low-Latency Security

Basic Multiplexing models. Computer Networks - Vassilis Tsaoussidis

Network Performance: Networks must be fast. What are the essential network performance metrics: bandwidth and latency

Introduction to PCI Express Positioning Information

EtherCAT Cutting Costs with High-speed Ethernet

Three Key Design Considerations of IP Video Surveillance Systems

Master Course Computer Networks IN2097

IT Data Communication and Networks (Optional)

MicroBlaze Debug Module (MDM) v3.2

Public Network. 1. Relatively long physical distance 2. Requiring a service provider (carrier) Branch Office. Home. Private Network.

(Refer Slide Time: 00:01:16 min)

Performance evaluation

PCI Express Impact on Storage Architectures and Future Data Centers. Ron Emerick, Oracle Corporation

Transcription:

Vodafone Chair Mobile Communications Systems, Prof. Dr.-Ing. G. Fettweis Interconnection Generation for System-on-Chip Design and Design Space Exploration Dipl.-Ing. Markus Winter Vodafone Chair for Mobile Communications Systems TU Dresden Kleinheubacher Tagung, 25. 26.09.2006 Integrierte digitale und analoge Schaltungen

TOMAHAWK System-on-Chip WIGWAM demonstration chip 1GBit/s data rate Proof of processor development concept (GenCore) Test of SoC design methodology Scope of this presentation: interconnection as critical part of system! TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 2

Agenda Introduction Problems at interconnection Idea of solution Implementation Results TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 3

Interconnection types Bus based system (tristate): Well-known and well-understood Easy to apply Little area Low protocol overhead Very slow High power consumption Circuit-switched star: Harder to apply (network logic) More area (multiplexer) Low protocol overhead Fast (parallel connections, no tristate) TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 4

Interconnection types Network-on-Chip: Packet-switched system Fast (parallel communication) GALS independent clock domains Complex protocol Error correction mechanisms Hard to apply (router development and linking) Problems with latency and packet loss depending on network load Much more area (multiple routers and FIFOs) TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 5

Problems Problems at application of interconnection scheme to system: What network architecture/topology shall be used? Who instantiates network/arbiter/routers and links s together? Who tests interconnection and communication? Reuse components developed once! Automate instantiation, linking, testing! TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 6

Generation flow Choose network type Define network parameters in XML Generate network description Instantiate and link network and s on top-level Processors with GenCore Instantiation, linking with network gen tool Fast SoC design/design space exploration TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 7

Network architecture Up to now: circuit-switched star topology in network generator Slave side arbitration parallel connections Easy protocol: 2 kinds of requests, 2 kinds of responses Master Network logic address request Slave address response address request write data request next address request read/write response TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 8

Network capabilities - throughput Maximum throughput: TP = n*f*b n number of maximum parallel connections; min(master, slave) f clock frequency B Bit width e.g. 4 masters, 5 slaves, 400MHz, 32 Bit data: TP = 51.2GBit/s Throughput in GBit/s 50 40 30 20 10 0 Theoratical maximum read transfers write transfers 5 10 15 20 Burst length TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 9

Network capabilities clock cycle time Minimum cycle time in ns Minimum cycle time depending on number of master 1.5 1 slave 4 slaves 1.25 6 slaves 10 slaves 1 1.75 0.5 2 4 6 8 10 Number of masters Time to compute arbitration, multiplexing, counter setting Indicates maximum possible clock frequency 600 MHz -> 1.6ns clock cycle, 0.8ns for signal propagation TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 10

Network capabilities area knand Area consumption depending on number of masters 20 1 slave 3 slaves 15 6 slaves 10 slaves 10 5 2 4 6 8 10 Number of masters Area for arbitration logic, multiplexing, counter setting Nearly no overhead -> scales completely with number of s TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 11

Thank you for your attention! TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 12

TU Dresden, 25.09.2006 Markus Winter, Gerhard Fettweis Slide 13