1 tuesday, september 26, 2006 wisdom consists of knowing when to avoid perfection. -horowitz

27
1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. - Horowitz

Post on 21-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

1

Tuesday, September 26, 2006

Wisdom consists of knowing when to avoid perfection.

- Horowitz

Page 2: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

2

Quiz 2Assignment 1

Page 3: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

3

Hypercube: log p dimensions with two nodes in each dimension

0-D hypercube

Page 4: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

4

Hypercube: log p dimensions with two nodes in each dimension

1-D hypercube

0-D hypercube

Page 5: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

5

Hypercube: log p dimensions with two nodes in each dimension

2-D hypercube

1-D hypercube

Page 6: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

6

Hypercube: log p dimensions with two nodes in each dimension

3-D hypercube

2-D hypercube

Page 7: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

7

Hypercube: log p dimensions with two nodes in each dimension

3-D hypercube

4-D hypercube

Each node is connected to d=log p other nodes

Page 8: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

8

•Numbering

•Minimum distance between nodes

Page 9: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

9

Diameter: Maximum distance between any two processing nodes in the network Ring 2-D Mesh Hypercube

Page 10: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

10

Diameter: Maximum distance between any two processing nodes in the network Ring

• └p/2┘ 2-D Mesh

• 2(√p -1) no-wraparound• 2 └(√p /2) ┘ wraparound

Hypercube• log p

Page 11: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

11

Connectivity: Multiplicity of paths Minimum arcs that need to be removed to disconnect

the network into twoRing

• 2 2-D Mesh

• 2 no-wraparound• 4 wraparound

Hypercube• d=log p

Page 12: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

12

Bisection width: Minimum arcs that need to be removed to partition the

network into two equal halvesRing

• 2 2-D Mesh

• √p no-wraparound• 2√p wraparound

Hypercube• p/2

Page 13: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

13

Page 14: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

14

Domain Decomposition

In this type of partitioning, the data associated with a problem is decomposed. Each parallel task then works on a portion of the data.

Page 15: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

15

Domain Decomposition

Page 16: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

16

Functional Decomposition

Page 17: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

17

Signal processing

Page 18: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

18

Climate modeling

.

Page 19: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

19

Examples of decomposition and task dependencies

Page 20: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

20

Examples of decomposition and task dependencies.

Page 21: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

21

Examples of decomposition and task dependencies.

Page 22: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

22

Granularity

Fine vs. Coarse Decomposition in large number of small tasks

vs. small number of large tasks.

Maximum degree of concurrencyAverage degree of concurrencyConcurrency vs. Granularity?

Page 23: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

23

Granularity

Page 24: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

24

Granularity

Critical Path length: Longest directed path between any pair of start

and finish nodes is critical path

Average degree of concurrency: Ratio of total amount of work to the critical

path length

Page 25: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

25

Granularity

•Another example

Page 26: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

26

Granularity

Measure of the ratio of computation to communication.

Fine-grain Parallelism: Facilitates load balancing Implies high communication overhead and less

opportunity for performance enhancement Coarse-grain Parallelism:

High computation to communication ratio Implies more opportunity for performance increase Harder to load balance efficiently

Page 27: 1 Tuesday, September 26, 2006 Wisdom consists of knowing when to avoid perfection. -Horowitz

27

Granularity

Example: Domain decompositions for a problem

involving a three-dimensional grid.