Parallel Processing & Distributed Systems Thoai Nam Chapter 3.

Slides:



Advertisements
Similar presentations
Efficient Realization of Hypercube Algorithms on Optical Arrays* Hong Shen Department of Computing & Maths Manchester Metropolitan University, UK ( Joint.
Advertisements

CIS December '99 Introduction to Parallel Architectures Dr. Laurence Boxer Niagara University.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Parallel Programming in C with MPI and OpenMP Michael J. Quinn.
1 Meshes of Trees (MoT) and Applications in Integer Arithmetic Panagiotis Voulgaris Petros Mol Course: Parallel Algorithms.
Jie Liu, Ph.D. Professor Department of Computer Science
Parallel Processing & Distributed Systems Thoai Nam Chapter 2.
Parallel Architectures: Topologies Heiko Schröder, 2003.
Parallel Architectures: Topologies Heiko Schröder, 2003.
1 Interconnection Networks Direct Indirect Shared Memory Distributed Memory (Message passing)
1 Lecture 23: Interconnection Networks Topics: communication latency, centralized and decentralized switches (Appendix E)
Advanced Topics in Algorithms and Data Structures An overview of the lecture 2 Models of parallel computation Characteristics of SIMD models Design issue.
Interconnection Networks 1 Interconnection Networks (Chapter 6) References: [1,Wilkenson and Allyn, Ch. 1] [2, Akl, Chapter 2] [3, Quinn, Chapter 2-3]
Parallel Routing Bruce, Chiu-Wing Sham. Overview Background Routing in parallel computers Routing in hypercube network –Bit-fixing routing algorithm –Randomized.
Chapter 9 Graph algorithms Lec 21 Dec 1, Sample Graph Problems Path problems. Connectedness problems. Spanning tree problems.
1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz.
Lecture 10 Outline Material from Chapter 2 Interconnection networks Processor arrays Multiprocessors Multicomputers Flynn’s taxonomy.
Models of Parallel Computation Advanced Algorithms & Data Structures Lecture Theme 12 Prof. Dr. Th. Ottmann Summer Semester 2006.
Interconnection Network Topologies
Examples Broadcasting and Gossiping. Broadcast on ring (Safe and Forward)
1 Lecture 24: Parallel Algorithms I Topics: sort and matrix algorithms.
Course Outline Introduction in algorithms and applications Parallel machines and architectures Overview of parallel machines, trends in top-500 Cluster.
1 Static Interconnection Networks CEG 4131 Computer Architecture III Miodrag Bolic.
ECE669 L16: Interconnection Topology March 30, 2004 ECE 669 Parallel Computer Architecture Lecture 16 Interconnection Topology.
Interconnection Networks. Applications of Interconnection Nets Interconnection networks are used everywhere! ◦ Supercomputers – connecting the processors.
Course Outline Introduction in software and applications. Parallel machines and architectures –Overview of parallel machines –Cluster computers (Myrinet)
Network Topologies Topology – how nodes are connected – where there is a wire between 2 nodes. Routing – the path a message takes to get from one node.
-1.1- PIPELINING 2 nd week. -2- Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM PIPELINING 2 nd week References Pipelining concepts The DLX.
CS668- Lecture 2 - Sept. 30 Today’s topics Parallel Architectures (Chapter 2) Memory Hierarchy Busses and Switched Networks Interconnection Network Topologies.
1 Interconnects Shared address space and message passing computers can be constructed by connecting processors and memory unit using a variety of interconnection.
CSE Advanced Computer Architecture Week-11 April 1, 2004 engr.smu.edu/~rewini/8383.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Parallel Programming in C with MPI and OpenMP Michael J. Quinn.
Minimal Fault Diameter for Highly Resilient Product Networks Khaled Day, Abdel-Elah Al-Ayyoub IEEE Trans. On Parallel and Distributed Systems 2000 vol.
Chapter 2 Parallel Architectures. Outline Interconnection networks Interconnection networks Processor arrays Processor arrays Multiprocessors Multiprocessors.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Parallel Programming in C with MPI and OpenMP Michael J. Quinn.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Parallel Programming in C with MPI and OpenMP Michael J. Quinn.
Lecture 3 Innerconnection Networks for Parallel Computers
Computer Science and Engineering Parallel and Distributed Processing CSE 8380 January Session 4.
-1- Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Parallel Computer Architectures 2 nd week References Flynn’s Taxonomy Classification of Parallel.
Course Outline Introduction in algorithms and applications Parallel machines and architectures Overview of parallel machines, trends in top-500, clusters,
Embedding long paths in k-ary n-cubes with faulty nodes and links
Lecture 3 TTH 03:30AM-04:45PM Dr. Jianjun Hu CSCE569 Parallel Computing University of South Carolina Department of.
Parallel Processing & Distributed Systems Thoai Nam And Vu Le Hung.
Complex Contagions Models in Opportunistic Mobile Social Networks Yunsheng Wang Dept. of Computer Science, Kettering University Jie Wu Dept. of Computer.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Shared versus Switched Media.
MATRIX MULTIPLICATION 4 th week. -2- Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM MATRIX MULTIPLICATION 4 th week References Sequential matrix.
1 Interconnection Networks. 2 Interconnection Networks Interconnection Network (for SIMD/MIMD) can be used for internal connections among: Processors,
Parallel Processing & Distributed Systems Thoai Nam Chapter 2.
Interconnection Networks
Basic Linear Algebra Subroutines (BLAS) – 3 levels of operations Memory hierarchy efficiently exploited by higher level BLAS BLASMemor y Refs. FlopsFlops/
Super computers Parallel Processing
Graph Theory. undirected graph node: a, b, c, d, e, f edge: (a, b), (a, c), (b, c), (b, e), (c, d), (c, f), (d, e), (d, f), (e, f) subgraph.
HYPERCUBE ALGORITHMS-1
Chapter 6 – Trees. Notice that in a tree, there is exactly one path from the root to each node.
COMP8330/7330/7336 Advanced Parallel and Distributed Computing Tree-Based Networks Cache Coherence Dr. Xiao Qin Auburn University
Interconnection Networks Communications Among Processors.
Distributed-Memory or Graph Models
Distributed and Parallel Processing
Computer Network Topology
Interconnection topologies
Course Outline Introduction in algorithms and applications
Computer Architecture Introduction to Data-Parallel architectures
Butterfly Network A butterfly network consists of (K+1)2^k nodes divided into K+1 Rows, or Ranks. Let node (i,j) refer to the jth node in the ith Rank.
Parallel Architectures Based on Parallel Computing, M. J. Quinn
Outline Interconnection networks Processor arrays Multiprocessors
Interconnection Network Design Lecture 14
Mesh-Connected Illiac Networks
Static Interconnection Networks
High Performance Computing & Bioinformatics Part 2 Dr. Imad Mahgoub
Interconnection Networks
Static Interconnection Networks
Presentation transcript:

Parallel Processing & Distributed Systems Thoai Nam Chapter 3

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Parallel Computer Architecture  Criteria: –Diameter, bisection width, etc.  Processor Organizations: –Mesh, binary tree, hypertree, pyramid, butterfly, hypercube, shuffle-exchange

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Criteria  Diameter The largest distance between two nodes Lower diameter is better  Bisection width The minimum number of edges that must be removed in order to divide the network into two halves (within one)  Number of edges per node  Maximum edge length

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Mesh (1)  Q-dimensional lattice  Communication is allowed only between neighboring nodes. Interior nodes communicate with 2q other nodes.

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Mesh (2)  Q-dimensional mesh with k q nodes –Diameter: q(k-1) –Bisection width: k q-1 –The maximum number of edges per node: 2q –The maximum edge length is a constant

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Binary Tree  Depth k-1: 2 k -1 nodes  Diameter: 2(k-1)  Bisection width: 1  Length of the longest edge: increasing

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Hypertree (1)  Hypertree of degree k and depth d: a complete k-ary tree of height d.

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Hypertree (2)  A 4-ary hypertree with depth d has 4 d leaves and 2 d (2 d+1 -1) nodes in all –Diameter: 2d –Bisection width: 2 d+1 –The number of edges per node  6 –Length of the longest edge: increasing

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Pyramid  Size k 2 : base a 2D mesh network containing k 2 processors, the total number of processors=(4/3)k 2 -1/3  A pyramid of size k 2 : –Diameter: 2logk –Bisection width: 2k –Maximum of links per node: 9 –Length of the longest edge: increasing

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Butterfly (1)  (k+1)2 k nodes divided into k+1 rows (rank), each contains n=2 k nodes.  Ranks are labeled 0 through k  Node(i,j): j-th node on the i-th rank  Node(i,j) is connected to two nodes on rank i-1: node(i-1,j) and node (i-1,m), where m is the integer found by inverting the i-th most significant bit in the binary representation of j  If node(i,j) is connected to node(i-1,m), then node (i,m) is connected to (i-1,j)  Diameter=2k  Bisection width=2 k-1  Length of the longest edge: increasing

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Butterfly (2) Rank 0 Rank 1 Rank 2 Rank 3 Node(1,5): i=1, j=5 j = 5 = 101 (binary) i=1 001 = 1 Node(1,5) is connected to node(0,1)

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Hypercube (1)  2 k nodes form a k-dimensional hypercube  Nodes are labeled 0, 1, 2,…, 2 k -1  Two nodes are adjacent if their labels differ in exactly one bit position  Diameter=k  Bisection width= 2 k -1  Number of edges per node is k  Length of the longest edge: increasing

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Hypercube (2)  5 = 0101  1 = 0001  4 = 0100  13 = 1101

Khoa Coâng Ngheä Thoâng Tin – Ñaïi Hoïc Baùch Khoa Tp.HCM Others  Cube-Connected cycles  Shuffle-Exchange  De Bruijn