1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 )...

36
1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭郭郭 ) 師師師師 師師師 2011/06/29 Joint work with F.H. Chang

Transcript of 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 )...

Page 1: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

1

Wide-Sense Nonblocking Under New Compound Routing Strategies

Junyi David Guo(郭君逸 )師範大學 數學系

2011/06/29

Joint work with F.H. Chang

Page 2: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

2

Applications

• Come from the need to interconnect telephones

• Interconnect processors with memories• Data transmission• Conference calls• Satellite communication

Page 3: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

3

One frequently discussed topic in switching networks is its nonblocking property.

Page 4: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

5

Multi-stage interconnection network

inputs outputs

stage

Page 5: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

6

Symmetry

• C(n, m, r)• C(2, 4, 3)

n 1

2

r

1

2

m

n1

2

r

Page 6: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

7

Definitions

• Request• Strictly nonblocking(SNB)• Wide-sense

nonblocking(WSNB)

Page 7: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

8

Matrix

1

2

4

2

O1 O2 O3

I1

I2

I3

1

3

Page 8: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

9

3-stage Clos Network

• [Clos 1953] C(n1, r1, m, n2, r2) is SNB iff m ≥ n1+n2-1.

• C(n, m, r) is SNB iff m ≥ 2n-1.

Page 9: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

10

[n+5,2n-2]

n+4

n+3

[12,n-1]

11[6,10]2n-14,51,2,3

n+2

n,n+1

Page 10: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

11

Algorithms

• Cyclic dynamic (CD)• Cyclic static (CS)• Save the unused (STU)• Packing (P)• Minimum index (MI)• Compound

Page 11: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

12

• Beneš(1965) proved that C(n, m, 2) is WSNB under P if and only if .

Page 12: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

13

Literature review

• Smith(1977) proved that C(n, m, r) is not WSNB under P or MI if .

• Du et al.(2001) improved to .• Hwang(2001) extend it to cover CS and CD.• Yang and Wang(1999) proved that C(n, m, r)

is not WSNB under P if .• Chang et al.(2004) improved to .• Chang et al.(2007) extend it to Multi-logd.

Page 13: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

16

Packing

• For P, hence STU, C(n, m, r), r ≥ 3, is WSNB iff m ≥ 2n-1.

[1,n]

[1,n]

n

[1,n-1]

n+1

n

[1,n-1]

n+1

n+1n

[1,n-1]

n+1

n+1n

[1,n-1]

n+1

n+1

n+2

n

[1,n-1]

n+1

n+2

n+2

n

[1,n-1]

n+1,n+2

n+2n

[1,n-1]

[n+1,2n-2]

Page 14: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

17

MI

• C(n, m, r) is WSNB under MI iff m ≥ 2n-1

X

Y

2n-3

Page 15: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

18

Packing+MI, STU+MI

• C(n, m, r), r ≥ 3, is WSNB iff m ≥ 2n-1.

Page 16: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

19

3-stage Clos Network

• C(n1, r1, m, n2, r2)

• C(2, 4, 3, 3, 2)

n1

n2

1

2

r1

r2

1

1

2

m

Page 17: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

20

Asymmetry

• C(n1, r1, m, n2, r2) is WSNB iff m ≥ n1+n2-1 under every known routing strategies.

Page 18: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

21

Multi-logd N Networks

• First proposed by Lea(1990).

Baseline

Baseline

Baseline

Page 19: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

23

Baseline Example

• BL2(4)

Page 20: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

26

Multi-logdN Network

• (Shyy & Lea 1991, Hwang 1998, Chang, Guo, and Hwang 2007) Multi-logdN network is SNB if and only if , where

odd. for 1d2

even, for 1)1()(

21-n

2 1

n

nddnp

n

• Maximal blocking configuration(MBC), is a set of requests blocking .

Page 21: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

28

Graph Model

• The graph model of BL2(4)

1

2

3

4

5

6

1

2

3

4

1

2

3

4

5

6

1

2

3

4

5

6

Page 22: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

29

P or STU

• Multi-logdN network is WSNB under P or STU iff p ≥ p(n).

Page 23: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

30

MI

• Multi-logdN network is WSNB under MI iff p ≥ p(n).

I1

I2

O1

O2

I1

I2

O1

O2

stage

2

1n

2

1n

2

2n

2

2n

0

21n

d

0

1

0

21n

d2

1n

d

0

21-n

d

21-n

d

21-n

d

21-n

d

Page 24: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

31

For n even

I1

I2

O1

O2

12

n

1I

2I

1O

2O

2

n1

2

n

0 0 0 0 0

12 n

d

12 n

d

2n

d

12)1( n

dd

12 2 n

d

1

1d

d

1d

12 d

1

12 n

d

12 n

d

12 n

d

12 n

d

2n

d

12 n

d

2n

d 2n

d

12)1( n

dd

12 2 n

d12 2

n

d

Page 25: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

32

P+MI and STU+MI

• Multi-logdN network is WSNB under P+MI and STU+MI iff p ≥ p(n).

Page 26: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

33

Generalizations

• General vertical-copy network.

Page 27: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

37

Open Problems

• Find a single-selected case for P, even STU, under 3-stage Clos network or multi-logd network.

• Multi-log network with extra stage.• MI under vertical-copy network• Find a general argument independent of

any routing algorithm. • Conjecture: There is no good WSNB

algorithm under one-to-one traffic.

Page 28: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

38

Thank you!

Page 29: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Fewest move

Alexander H. Frey, Jr. and David Singmaster. Handbook of Cubik Math. Enslow Publishers, 1982.◦最少 17步,最多 52步◦猜測 : 「上帝的數字 (God’s number)」為 20。

43

任意的魔術方塊,在幾步內可以完成呢?

Page 30: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Fewest move

Michael Reid.◦ Lower bounds: Superflip requires 20 face turns. (1995)

U R2 F B R B2 R U2 L B2 R U' D' R2 F R' L B2 U2 F2

◦Upper bounds: New upper bounds. (1995) Face turn metric(FTM): 29 moves. Quarter turn metric(QTM): 42 moves.

44

Page 31: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Fewest move

Silviu Radu.◦Rubik can be solved in 27f. (2006/04)◦Paper. (June 30, 2007)

Face turn metric(FTM): 27 moves. Quarter turn metric(QTM): 34 moves.

◦Using GAP(Groups, Algorithms, Programming):a System for Computational Discrete Algebra

45

Page 32: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Fewest moveDaniel Kunkle and Gene Cooperman, 26

Moves Suffice for Rubik's Cube, Proc. of International Symposium on Symbolic and Algebraic Computation (ISSAC '07), ACM Press, 2007.◦FTM: 26 moves.◦Gene Cooperman, Larry Finkelstein, and Namita

Sarawagi. Applications of Cayleygraphs. In AAECC: Applied Algebra, Algebraic Algorithmsand Error-Correcting Codes, InternationalConference, pages 367-378 LNCS, Springer-Verlag, 1990. FTM: 11 moves. QTM: 14 moves. 46

Page 33: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Fewest move

Tomas Rokicki.◦25 Moves Suffice for Rubik’s Cube. (2008.3)◦23 Moves Suffice for Rubik’s Cube. (2008.4)◦22 Moves Suffice for Rubik’s Cube. (2008.8)◦God’s Number is 20. (2010.7)

47

Page 34: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

3x3 方塊的變化數There are

Cube subgroups.

48

( 8 ! ∙ 38

3 )×( 12 ! ∙ 212

2 )×12=43252003274489856000

¿ 4.3 ×1019

Page 35: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

Partition into 2,217,093,120 sets of 19,508,428,800 positions.

Reduce to 55,882,296.About 35 CPU years.

49

God’s Number is 20

Page 36: 1 Wide-Sense Nonblocking Under New Compound Routing Strategies Junyi David Guo( 郭君逸 ) 師範大學 數學系 2011/06/29 Joint work with F.H. Chang.

50

Distance Count of Positions

0 1

1 18

2 243

3 3,240

4 43,239

5 574,908

6 7,618,438

7 100,803,036

8 1,332,343,288

9 17,596,479,795

10 232,248,063,316

11 3,063,288,809,012

12 40,374,425,656,248

13 531,653,418,284,628

14 6,989,320,578,825,358

15 91,365,146,187,124,313

16 about 1,100,000,000,000,000,000

17 about 12,000,000,000,000,000,000

18 about 29,000,000,000,000,000,000

19 about 1,500,000,000,000,000,000

20 about 300,000,000

Number of Positions