Effective bandwidth with link pipelining Pipeline the flight and transmission of packets over the links Overlap the sending overhead with the transport.

Slides:



Advertisements
Similar presentations
Interconnection Networks: Flow Control and Microarchitecture.
Advertisements

A Novel 3D Layer-Multiplexed On-Chip Network
Presentation of Designing Efficient Irregular Networks for Heterogeneous Systems-on-Chip by Christian Neeb and Norbert Wehn and Workload Driven Synthesis.
Fundamentals of Computer Networks ECE 478/578 Lecture #13: Packet Switching (2) Instructor: Loukas Lazos Dept of Electrical and Computer Engineering University.
Flattened Butterfly Topology for On-Chip Networks John Kim, James Balfour, and William J. Dally Presented by Jun Pang.
Interconnection Networks: Topology and Routing Natalie EnrightJerger.
1 Message passing architectures and routing CEG 4131 Computer Architecture III Miodrag Bolic Material for these slides is taken from the book: W. Dally,
Flattened Butterfly: A Cost-Efficient Topology for High-Radix Networks ______________________________ John Kim, William J. Dally &Dennis Abts Presented.
High Performance Router Architectures for Network- based Computing By Dr. Timothy Mark Pinkston University of South California Computer Engineering Division.
1 Lecture 12: Interconnection Networks Topics: dimension/arity, routing, deadlock, flow control.
1 Lecture 23: Interconnection Networks Topics: communication latency, centralized and decentralized switches (Appendix E)
ECE669 L12: Interconnection Network Performance March 9, 2004 ECE 669 Parallel Computer Architecture Lecture 12 Interconnection Network Performance.
NUMA Mult. CSE 471 Aut 011 Interconnection Networks for Multiprocessors Buses have limitations for scalability: –Physical (number of devices that can be.
Computer Networks: Performance Measures1 Computer Network Performance Measures.
1 Lecture 24: Interconnection Networks Topics: communication latency, centralized and decentralized switches (Sections 8.1 – 8.5)
Issues in System-Level Direct Networks Jason D. Bakos.
Networks: Performance Measures1 Network Performance Measures.
1 Lecture 24: Interconnection Networks Topics: topologies, routing, deadlocks, flow control.
1 Indirect Adaptive Routing on Large Scale Interconnection Networks Nan Jiang, William J. Dally Computer System Laboratory Stanford University John Kim.
1 Lecture 25: Interconnection Networks Topics: communication latency, centralized and decentralized switches, routing, deadlocks (Appendix E) Review session,
1 Static Interconnection Networks CEG 4131 Computer Architecture III Miodrag Bolic.
Performance and Power Efficient On-Chip Communication Using Adaptive Virtual Point-to-Point Connections M. Modarressi, H. Sarbazi-Azad, and A. Tavakkol.
Switching, routing, and flow control in interconnection networks.
Interconnect Network Topologies
Interconnection Networks. Applications of Interconnection Nets Interconnection networks are used everywhere! ◦ Supercomputers – connecting the processors.
High Performance Embedded Computing © 2007 Elsevier Lecture 16: Interconnection Networks Embedded Computing Systems Mikko Lipasti, adapted from M. Schulte.
1 Lecture 23: Interconnection Networks Topics: Router microarchitecture, topologies Final exam next Tuesday: same rules as the first midterm Next semester:
1 The Turn Model for Adaptive Routing. 2 Summary Introduction to Direct Networks. Deadlocks in Wormhole Routing. System Model. Partially Adaptive Routing.
Distributed Quality-of-Service Routing of Best Constrained Shortest Paths. Abdelhamid MELLOUK, Said HOCEINI, Farid BAGUENINE, Mustapha CHEURFA Computers.
Interconnect Networks
On-Chip Networks and Testing
Introduction to Interconnection Networks. Introduction to Interconnection network Digital systems(DS) are pervasive in modern society. Digital computers.
Distributed Routing Algorithms. In a message passing distributed system, message passing is the only means of interprocessor communication. Unicast, Multicast,
1 Interconnects Shared address space and message passing computers can be constructed by connecting processors and memory unit using a variety of interconnection.
High-Level Interconnect Architectures for FPGAs An investigation into network-based interconnect systems for existing and future FPGA architectures Nick.
Improving Capacity and Flexibility of Wireless Mesh Networks by Interface Switching Yunxia Feng, Minglu Li and Min-You Wu Presented by: Yunxia Feng Dept.
High-Level Interconnect Architectures for FPGAs Nick Barrow-Williams.
Multiprocessor Interconnection Networks Todd C. Mowry CS 740 November 3, 2000 Topics Network design issues Network Topology.
1 Message passing architectures and routing CEG 4131 Computer Architecture III Miodrag Bolic Material for these slides is taken from the book: W. Dally,
Network-on-Chip Introduction Axel Jantsch / Ingo Sander
Anshul Kumar, CSE IITD CSL718 : Multiprocessors Interconnection Mechanisms Performance Models 20 th April, 2006.
CS 8501 Networks-on-Chip (NoCs) Lukasz Szafaryn 15 FEB 10.
InterConnection Network Topologies to Minimize graph diameter: Low Diameter Regular graphs and Physical Wire Length Constrained networks Nilesh Choudhury.
Anshul Kumar, CSE IITD ECE729 : Advanced Computer Architecture Lecture 27, 28: Interconnection Mechanisms In Multiprocessors 29 th, 31 st March, 2010.
Computer Science and Engineering Copyright by Hesham El-Rewini Advanced Computer Architecture.
Interconnect Networks Basics. Generic parallel/distributed system architecture On-chip interconnects (manycore processor) Off-chip interconnects (clusters.
LECTURE 12 NET301 11/19/2015Lect NETWORK PERFORMANCE measures of service quality of a telecommunications product as seen by the customer Can.
Networks: Routing, Deadlock, Flow Control, Switch Design, Case Studies Alvin R. Lebeck CPS 220.
Topology How the components are connected. Properties Diameter Nodal degree Bisection bandwidth A good topology: small diameter, small nodal degree, large.
1 Lecture 24: Interconnection Networks Topics: communication latency, centralized and decentralized switches, routing, deadlocks (Appendix F)
1 Lecture 14: Interconnection Networks Topics: dimension vs. arity, deadlock.
Interconnection Networks Communications Among Processors.
Review of Useful Definitions Statistical multiplexing is a method of sharing a link among transmissions. When computers use store-and-forward packet switching,
1 Lecture 22: Interconnection Networks Topics: Routing, deadlock, flow control, virtual channels.
COMP8330/7330/7336 Advanced Parallel and Distributed Computing Communication Costs in Parallel Machines Dr. Xiao Qin Auburn University
Chapter 3 Part 3 Switching and Bridging
Lecture 23: Interconnection Networks
Interconnection topologies
Azeddien M. Sllame, Amani Hasan Abdelkader
Chapter 3 Part 3 Switching and Bridging
Computer Network Performance Measures
CprE 458/558: Real-Time Systems
Introduction to Scalable Interconnection Networks
Lecture: Interconnection Networks
Computer Network Performance Measures
Network Topologies Charles Warren.
Embedded Computer Architecture 5SAI0 Interconnection Networks
Chapter 3 Part 3 Switching and Bridging
Circuit Switched Network
Chapter 2 from ``Introduction to Parallel Computing'',
Presentation transcript:

Effective bandwidth with link pipelining Pipeline the flight and transmission of packets over the links Overlap the sending overhead with the transport latency and receiving overhead of prior packets Sending overhead Transport latency time overlap Receiving overhead

Injection bandwidth Network injection Reception bandwidth Network reception Aggregate bandwidth Characterizing Performance : Effective Bandwidth Eff. bandwidth = min (BW NetworkInjection, BW NetworkReception ) = min (NxBW LinkInjection, NxBW LinkReception ) = min (2xBW LinkInjection, 2xBW LinkReception )

BW LinkInjection = Packet size max (sending overhead, transmission time) BW LinkReception = Packet size max (receiving overhead, transmission time) Eff. bandwidth = min (NxBW LinkInjection, NxBW LinkReception ) = N x Packet size max (overhead, transmission time) overhead = max (sending overhead, receiving overhead)

Characterizing Performance: Effective Bandwidth A Simple (General) Throughput Performance Model: The network can be considered as a “pipe” of variable width There are three points of interest end-to-end: –Injection into the pipe –Narrowest section within pipe (i.e., minimum network bisection that has traffic crossing it) –Reception from the pipe Injection bandwidth Bisection bandwidth Reception bandwidth

Effective bandwidth = min(BW NetworkInjection, BW Network, σ × BW NetworkReception ) = min(N × BW LinkInjection, BW Network, σ × N × BW LinkReception ) BW Network = ρ × BW Bisection 

BW Network = ρ × BW Bisection × 8/3 Characterizing Performance: Effective Bandwidth Injection bandwidth Network injection Reception bandwidth Network reception Aggregate bandwidth unidirectional ring greedy traffic: node i sends to node i + 3 mod N Bisection Bandwidth   = 3/8

Simple (General) Model Applied to Interconnecting Two Devices: Effective bandwidth = min(2 × BW LinkInjection, BW Network, 1 × (2 × BW LinkReception )) BW Network =  L × 2 × BW Link 1 Dedicated-link network int. network Device A Device B  L = link efficiency resulting from flow control, encoding, packet header and trailer overheads BW Link

2D torus of 16 nodes hypercube of 16 nodes (16 = 2 4, so n = 4) 2D mesh or grid of 16 nodes Network Bisection

An NoC architecture can be uniquely described by the triple Arch(T(R,Ch), P R, Ω(C)), where, 1.The labeled graph T(R,Ch) represents the network topology. The routers and channels in the network are given by the sets R and Ch, respectively 2.{P R (r, i, j)|i, j, r ∈ R} defines the routing policy P R at router r, for any source router i and destination router j, while considering a particular switching technique. 3.Ω : C → R is a function that maps each vertex c i ∈ C in the APCG to a router in R. Network-on-Chip

On-chip-network building block Topology  The on-chip-network topology determines the physical layout and connection between nodes and channel in the network.  Metrics for comparing topology  Degree  Hop Count  Maximum channel node  Path Diversity

Example:

Routing  The routing algorithm is used to decide what path a message will take through the network to reach its destination.  Types of routing algorithm  Deterministic routing algorithm  Oblivious routing algorithm  Adaptive routing algorithm

Deterministic routing algorithm m i = d i − s i mod k Δ i = m i – 0 if m i ≤ k/2 otherwise m i - k This can then be used to compute our preferred directions: D T,i = 0 if |Δ i | = k/2 otherwise sign(Δ i ) Deterministic routing algorithm send every packet from source x to destination y over exactly the same route.

Oblivious routing algorithm Oblivious routing, in which we route packets without regard for the state of the network, is simple to implement and simple to analyze. Intermediate node

Minimal Quadrant Possible routes

Adaptive routing algorithm An adaptive routing algorithm uses information about the network state, typically queue occupancies, to select among alternative paths to deliver a packet. Partial Adaptive routing algorithm

a. Fully Adaptive routing algorithm b. Deadlock condition