Meshes of Trees (MoT) and Applications in Integer Arithmetic - - PowerPoint PPT Presentation

meshes of trees mot and applications in integer arithmetic
SMART_READER_LITE
LIVE PREVIEW

Meshes of Trees (MoT) and Applications in Integer Arithmetic - - PowerPoint PPT Presentation

Meshes of Trees (MoT) and Applications in Integer Arithmetic Panagiotis Voulgaris Petros Mol Course: Parallel Algorithms 1 Outline of the talk The two-Dimensional Mesh of Trees Definitions Properties Variations Integer


slide-1
SLIDE 1

1

Meshes of Trees (MoT) and Applications in Integer Arithmetic

Panagiotis Voulgaris Petros Mol Course: Parallel Algorithms

slide-2
SLIDE 2

2

Outline of the talk

The two-Dimensional Mesh of Trees

  • Definitions
  • Properties
  • Variations

Integer Arithmetic Applications

  • Multiplication
  • Division
  • Powering
  • Root Finding
slide-3
SLIDE 3

3

Definition

Construction:

N N ×

grid Mesh of Trees

2

3 2 N N −

Nodes

slide-4
SLIDE 4

4

Properties

1)Diameter (maximum distance between any pair of processors): 4logN

Proof u v u v Case 1: u belongs to a row tree and v to a column tree Dist<=2logN +2logN

slide-5
SLIDE 5

5

Properties (cont.)

Case 2: u,v belong only to row trees (or only to column trees)

u v u u u v u u Dist=logN –r +2logN + logN + s<=4logN since r>=s

slide-6
SLIDE 6

6

Properties (cont)

Thus meshes of trees enjoy both small diameter and large bisection width. This fact makes them a more efficient structure than arrays and simple trees

2)Bisection Width( the minimum number of wires that have to be removed in order to disconnect the network into two halves with “almost” identical number of processors) : N (Proof omitted)

slide-7
SLIDE 7

7

Recursive Decomposition

N N ×

Mesh of trees

2 2 N N ×

Four disjoint copies of Mesh of trees Importance: This property makes mesh of trees appropriate for recursive algorithms for parallel computation

slide-8
SLIDE 8

8

“Ideal” Parallel Computer

P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P+M P: Processor M: Memory

Every processor is linked to every

  • ther processor.

Advantage: Speed !! Drawback: Cost

slide-9
SLIDE 9

9

“Ideal” Parallel Computer

P P P P P P P P M M M M M M M M

Process/Memory separation

Again here the degree of each node becomes large as the number of processor increases Idea: Why not “simulate” the complete bipartite graph? Every Processor has direct access to a memory register

slide-10
SLIDE 10

10

“Ideal” Parallel Computer

slide-11
SLIDE 11

11

“Ideal” Parallel Computer

slide-12
SLIDE 12

12

“Ideal” Parallel Computer

slide-13
SLIDE 13

13

Benefits and Drawbacks

+ Simulation of any step of in 2logN steps

, N N

K

, N N

K + Bounded degree graph with essentially the

computational power as

+ We have actually constructed the NxN mesh of

Trees

  • The mesh of Trees has nearly nodes

whereas the initial complete bipartite graph had

  • nly 2N

Solution: Later

2

3N

slide-14
SLIDE 14

14

Back

Transformation to mesh of Trees

slide-15
SLIDE 15

15

Variations

1)

slide-16
SLIDE 16

16

Variations (cont)

slide-17
SLIDE 17

17

Variations (cont)

2)

slide-18
SLIDE 18

18

Variations (cont.)

3)

slide-19
SLIDE 19

19

Variations (cont.)

4)