Download presentation
Presentation is loading. Please wait.
Published byGeorge Simmons Modified over 9 years ago
1
Renato Hentschke Jaganathan Narasimham Marcelo Johann Ricardo Reis Renato Hentschke Jaganathan Narasimham Marcelo Johann Ricardo Reis IBM - Thomas Watson RC jagan@us.ibm.com Universidade Federal do Rio Grande do Sul {renato,johann,reis}@inf.ufrgs.br ISPD 2007 Maze Routing Steiner Trees With Delay vs. Wire Length Trade-off
2
Where are the rest of us from? Porto Alegre RS - Brasil
3
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US This work presents… AMAZE This work presents… AMAZE Let us start with a little… FAQ
4
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US What is AMAZE? What is AMAZE? An algorithm for single net Steiner Tree construction that uses A* searches and trades-off WL for delay Can do both Actual Routing Tree Topologies
5
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US In which context AMAZE fits in? In which context AMAZE fits in? Steiner Trees are used for estimation and perhaps global routing Maze Router is widely used for actual routing (global or detailed) Somewhere in the flow we decompose the nets by setting Steiner points or just forget our Steiner Trees and let the router run…
6
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US What is the motivation? What is the motivation? Maze Router has a lot of potential: Net ordering can be aliviated by negotiation Speed can be improved (a lot) with A* searches and by tuning data structures Memory is not a concern for most applications Degrees of freedom can be exploited to get quality results, e.g., good Trees Keep Maze router’s nice properties
7
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US How good can AMAZE be? How good can AMAZE be? What about: Within 2% from optimal Steiner Trees? From 26% to 40% improvement in delay compared to AHHK? From 1% to 30% improvement in delay compared to P-Trees? As fast as any heuristic Steiner Tree algorithm
8
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Is AMAZE a new algorithm at all? Is AMAZE a new algorithm at all? Hannan Grid Multiple src tgts Our Biasing Path Length factor Sharing factor Standard techniques: New Contributions Data Structures AMAZE A* search
9
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US What not to expect from AMAZE? What not to expect from AMAZE? The main weaknesses are (currently): Cannot guarantee efficiency (speed) or quality (good trees) with costs that vary too much. Problem for global routing that has to model congestion We have not checked if early estimation matches actual routing if context changes
10
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Outline Outline 1.Delay and WL trade-off 2.Properties of A* 3.Creating Steiner Trees with A* 4.Improving Wirelenght 5.Improving Delay to critical sinks 6.Experimental results 7.Concluding remarks
11
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 1. Delay and WL trade-off MRSTMRSA BRSTStarCSA
12
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 2. Basics of A* Expand more promising nodes first, acording to: f(n) = g(n) + h(n) origingoal Intermediate searched space Complete Search g(n)h(n) Effect of better estimation (higher h)
13
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g= h= f = 1 g=0 h=3 f =3 7 g= h= f = 13 g= h= f = 2 g= h= f = 8 g= h= f = 14 g= h= f = 9 g= h= f = 6 Open List: 7 c=1c=2 c=1 g= h= f = 3 c=1c=2g= h= f = 15 c=1c=2g= h= f = 10 c=2g= h= f = 4 c=1c=2g= h= f = 11 c=1c=2 Non uniform costs
14
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g= h= f = 9 Open List: 8, 1, 13, 6 c=2g= h= f = 3 c=1c=2g= h= f = 15 c=1c=2g= h= f = 10 c=2g= h= f = 4 c=1c=2g= h= f = 11 c=1c=2g=1 h=4 f =5 1 g=0 h=3 f =3 7 g=1 h=4 f =5 13 g= h= f = 2 g=2 h=2 f =4 8 g= h= f = 14 g=1 h=5 f =5 6 c=1c=2c=1 Non uniform costs
15
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 1, 13, 6, 9, 2, 14 g= h= f = 10 c=2g= h= f = 4 c=1c=2g= h= f = 11 c=1c=2g=1 h=4 f =5 1 g=0 h=3 f =3 7 g=1 h=4 f =5 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=3 h=3 f =6 14 g=4 h=1 f =5 9 g=1 h=4 f =5 6 c=1c=2 c=1 g= h= f = 3 c=1c=2g= h= f = 15 c=1c=2 Non uniform costs
16
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g= h= f = 10 c=2g= h= f = 4 c=1c=2g= h= f = 11 c=1c=2g=4 h=1 f =5 9 c=2g= h= f = 3 c=1c=2g= h= f = 15 c=1c=2g=1 h=4 f =5 1 7 g=1 h=4 f =5 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=3 h=3 f =6 14 g=1 h=4 f =5 6 c=1c=2c=1 g=0 h=3 f =3 Open List: 1, 13, 6, 9, 2, 14Ties Non uniform costs
17
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g= h= f = 10 c=2g= h= f = 4 c=1c=2g= h= f = 11 c=1c=2g=4 h=1 f =5 9 c=2g= h= f = 3 c=1c=2g= h= f = 15 c=1c=2g=1 h=4 f =5 1 7 g=1 h=4 f =5 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=3 h=3 f =6 14 g=1 h=4 f =5 6 c=1c=2c=1 g=0 h=3 f =3 Open List: 1, 13, 6, 9, 2, 14 But not critical as f * = 6 Non uniform costs
18
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 13 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g= h= f = 1 g= h= f = 7 g=0 h=4 f =4 13 g= h= f = 2 g= h= f = 8 g= h= f = 14 g= h= f = 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g= h= f = 15 c=1 Uniform costs
19
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 14,7 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g= h= f = 2 g= h= f = 8 g=1 h=3 f =4 14 g= h= f = 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g= h= f = 15 c=1 Critical Ties f(14) = f(7) = f *
20
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 15,8,7 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g= h= f = 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g= h= f = 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1 Can choose what you want
21
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 11,9,8,7 g= h= f = 10 c=1g= h= f = 4 c=1 g=3 h=1 f =4 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g=3 h=1 f =4 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1 Nodes close to the target first worst best
22
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 15,7,2,3 Nodes close to the target first Only four nodes expanded g=4 h=0 f =4 10 c=1g= h= f = 4 c=1 g=3 h=1 f =4 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g=3 h=1 f =4 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1
23
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g=4 h=0 f =4 10 c=1g= h= f = 4 c=1 g=3 h=1 f =4 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g=3 h=1 f =4 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1 Open List: 15,7,2,3 But pay attention to: Degree of freedom
24
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g=4 h=0 f =4 10 c=1g= h= f = 4 c=1 g=3 h=1 f =4 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g=3 h=1 f =4 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1 Open List: 15,7,2,3 But pay attention to: Degree of freedom
25
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US g=4 h=0 f =4 10 c=1g= h= f = 4 c=1 g=3 h=1 f =4 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g=3 h=3 f =6 2 g=2 h=2 f =4 8 g=1 h=3 f =4 14 g=3 h=1 f =4 9 g= h= f = 6 c=1 g= h= f = 3 c=1 g=2 h=2 f =4 15 c=1 Open List: 15,7,2,3 But pay attention to: Degree of freedom
26
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 15,7 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g= h= f = 9 g= h= f = 6 c=1c=2c=1 c=2 g= h= f = 3 c=1g=2 h=2 f =4 15 c=1 With Hannan Grid Freedom was lost… unless
27
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 15,7 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g= h= f = 1 g=1 h=3 f =4 7 g=0 h=4 f =4 13 g= h= f = 9 g= h= f = 6 c=1c=2c=1 c=2 g= h= f = 3 c=1g=2 h=2 f =4 15 c=1 With Hannan Grid We check number of steps
28
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Summary Simple ties f = fall expanded Critical ties f=f=f*break for efficiency Depth ties f=f=f* and g=gtrue freedom Stepped depth ties f=f=f* and s=s w/Hannan Net result: Best efficiency, freedom to chose the paths
29
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 3. Creating Steiner Trees with A* Sinks are added one by one Multiple Sourcesconnected tree Which target to connect next? Multiple Targetsselect the best accounts for blockages h function sees all sinks chosen target ct(n) used for speed First route all critical sinks (later on)
30
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: h Heuristic attracts search to closest node Formation of Trees
31
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
32
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
33
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
34
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
35
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
36
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Sources: Targets: Formation of Trees
37
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Multi-pin nets Multiple sources and targets Routing is complete for this simple case… But other pin dispositions require expertise Formation of Trees
38
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Select targets with A* Sinks are added one by one Multiple Sourcesconnected tree Which target to connect next? Multiple Targetsselect the best accounts for blockages h function sees all sinks chosen target ct(n) used for speed First route all critical sinks (later on)
39
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Naturally handle blockages
40
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 4. Improving Wirelenght Share as many segments as possible Use decision points (freedom) to chose paths closer to other pins Performed by the biasing funcion called when stepped depth ties happen
41
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Biasing computation Improves WL by sharing as many paths as possible 1. Eliminate nodes behind target and parent node 2. And closer to the tree 3. Compute centroid 4. Take the closest node
42
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Effect of biasing AMAZE with biasing off AMAZE with biasing ON
43
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 5. Improving Delay to critical sinks Route them first Use repulsive biasing Sharing Factor Path Length Factor Use as little sharing as possible
44
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Sharing factor Only for critical
45
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Path lenght factor Makes shortest paths from the source to the sinks Only for critical
46
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 6. Experimental results Average of 100 randomly generated nets
47
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Experimental results Average of 100 randomly generated nets
48
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Sample Trees
49
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Blockage Analysis
50
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Sample Trees P-TreesAMAZE
51
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Trade-off ranges
52
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 7. Concluding remarks Steiner Trees within 2% from optimal Savings of up to 40% in delay / AHHK Savings of up to 30% in delay / P-Trees Wider trade-off ranges for delay and WL Handle blockages better than competitors Flexibility and Speed Application: play with factors and get the best
53
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Questions? downtown Porto Alegre
54
Maze Routing Steiner Trees With Delay vs. Wire Length Trade-off Renato Hentschke Jaganathan Narasimham Marcelo Johann Ricardo Reis Renato Hentschke Jaganathan Narasimham Marcelo Johann Ricardo Reis IBM - Thomas Watson RC jagan@us.ibm.com Universidade Federal do Rio Grande do Sul {renato,johann,reis}@inf.ufrgs.br ISPD 2007 Thank you!
55
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US 3.2 Princípios da Pesquisa 3.2 Princípios da Pesquisa A partir de s, formar uma árvore de pesquisa pela aplicação repetitiva do operador de sucessão Um nodo é expandido quando se aplica a operação de sucessão sobre ele (o nodo se torna fechado) Um nodo é gerado quando é retornado pela operação de sucessão (o nodo se torna aberto) v1 v2 v3 v4 v5 v6 v7 v8 v9 s t v0 v5 v1 v6 v3 v8 v7 v9 v2 v4 admissibility
56
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Notação Notação s, t - source and target nodes k(n1,n2) - estimativa entre n1 e n2 k*(n1,n2) - custo real entre n1 e n2 h(n) = k(n,t) h*(n) = k*(n,t) g*(n) = k*(s,n) P a-b = caminho de a até b P a-b * = caminho ótimo de a até b f*(n) = custo de P s-n * P n-t * (passando por n)
57
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Propriedades em pesquisa heurística Propriedades em pesquisa heurística *Admissibilidade (*): custo de n a t h(n) garante menor caminho Consistência: k(n1,n2) + k(n2,n3) k(n1,n3) só expande nodos com custo mínimo conhecido: n1n1 n2n2 n3n3 k(n 1,n 2 ) k(n 2,n 3 ) k(n 1,n 3 ) t n h(n) Menor caminho de n a t g(n) = g*(n)
58
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Propriedades do Algoritmo A* Propriedades do Algoritmo A* Continuidade - sempre existe nodo P s-t *. Perfeição - A* não entra em loop. Admissibilidade - A* encontra caminho ótimo. Condição suficiente para expansão - f(n) < P s-t *. Condição necessária para expansão - f(n) P s-t *. Indeterminismo em empates críticos - biasing. Monotonia - se n 2 é expandido depois de n 1 então f(n 2 ) f(n 1 ). Dominância - A* mais informado expande menos. Excelência - A* ótimo entre unidirecionais informados quando heurística é consistente e sem empates críticos.
59
Hentschke, Narasimhan, Johann, Reis - ISPD 2007 - Austin, TX - US Open List: 13 g= h= f = 10 c=1g= h= f = 4 c=1 g= h= f = 11 c=1 g=1 h=4 f =5 1 g=0 h=3 f =3 7 g=0 h=4 f =4 13 g=2 h=3 f =5 2 g=1 h=2 f =3 8 g=3 h=3 f =6 14 g=2 h=1 f =3 9 g=1 h=4 f =5 6 c=1 g= h= f = 3 c=1 g= h= f = 15 c=1
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.