Multilevel Hypergraph Partitioning G. Karypis, R. Aggarwal, V. Kumar, and S. Shekhar Computer Science Department, U of MN Applications in VLSI Domain.

Slides:



Advertisements
Similar presentations
Multilevel Hypergraph Partitioning Daniel Salce Matthew Zobel.
Advertisements

Resource Management §A resource can be a logical, such as a shared file, or physical, such as a CPU (a node of the distributed system). One of the functions.
METIS Three Phases Coarsening Partitioning Uncoarsening
VLSI Layout Algorithms CSE 6404 A 46 B 65 C 11 D 56 E 23 F 8 H 37 G 19 I 12J 14 K 27 X=(AB*CD)+ (A+D)+(A(B+C)) Y = (A(B+C)+AC+ D+A(BC+D)) Dr. Md. Saidur.
Faster SAT and Smaller BDDs via Common Function Structure Fadi A. Aloul, Igor L. Markov, Karem A. Sakallah University of Michigan.
Parallel Simulation etc Roger Curry Presentation on Load Balancing.
FAST: A Novel Protein Structure Alignment Algorithm Jianhua Zhu and Zhiping Weng PROTEINS: Structure, Function, and Bioinformatics 58:618–627 (2005) Created.
Chapter 2 – Netlist and System Partitioning
A High Performance Application Representation for Reconfigurable Systems Wenrui GongGang WangRyan Kastner Department of Electrical and Computer Engineering.
Graph Coloring Algorithms for Fast Evaluation of Curtis Decomposition Marek Perkowski, Rahul Malvi, Stan Grygiel, Mike Burns, and Alan Mishchenko Portland.
A scalable multilevel algorithm for community structure detection
SubSea: An Efficient Heuristic Algorithm for Subgraph Isomorphism Vladimir Lipets Ben-Gurion University of the Negev Joint work with Prof. Ehud Gudes.
Can Recursive Bisection Alone Produce Routable Placements? Andrew E. Caldwell Andrew B. Kahng Igor L. Markov Supported by Cadence.
1 Circuit Partitioning Presented by Jill. 2 Outline Introduction Cut-size driven circuit partitioning Multi-objective circuit partitioning Our approach.
15-853Page :Algorithms in the Real World Separators – Introduction – Applications.
Multilevel Graph Partitioning and Fiduccia-Mattheyses
Joanna Ellis-Monaghan, St. Michaels College Paul Gutwin, Principal Technical Account Manager, Cadence.
VLSI Physical Design: From Graph Partitioning to Timing Closure Chapter 5: Global Routing © KLMH Lienig 1 FLUTE: Fast Lookup Table Based RSMT Algorithm.
Domain decomposition in parallel computing Ashok Srinivasan Florida State University COT 5410 – Spring 2004.
Graph clustering Jin Chen CSE Fall 2012 MSU 1.
Graph partition in PCB and VLSI physical synthesis Lin Zhong ELEC424, Fall 2010.
CHAMELEON : A Hierarchical Clustering Algorithm Using Dynamic Modeling
Escape Routing For Dense Pin Clusters In Integrated Circuits Mustafa Ozdal, Design Automation Conference, 2007 Mustafa Ozdal, IEEE Trans. on CAD, 2009.
A Fast Algorithm for Enumerating Bipartite Perfect Matchings Takeaki Uno (National Institute of Informatics, JAPAN)
Graph Partitioning Donald Nguyen October 24, 2011.
Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear.
Architectural Support for Fine-Grained Parallelism on Multi-core Architectures Sanjeev Kumar, Corporate Technology Group, Intel Corporation Christopher.
UNIVERSITAT POLITÈCNICA DE CATALUNYA Departament d’Arquitectura de Computadors Exploiting Pseudo-schedules to Guide Data Dependence Graph Partitioning.
March 20, 2007 ISPD An Effective Clustering Algorithm for Mixed-size Placement Jianhua Li, Laleh Behjat, and Jie Huang Jianhua Li, Laleh Behjat,
Parallel Computing Sciences Department MOV’01 Multilevel Combinatorial Methods in Scientific Computing Bruce Hendrickson Sandia National Laboratories Parallel.
Graph Coloring. Introduction When a map is colored, two regions with a common border are customarily assigned different colors. We want to use a small.
CS 584. Load Balancing Goal: All processors working all the time Efficiency of 1 Distribute the load (work) to meet the goal Two types of load balancing.
Efficient Route Computation on Road Networks Based on Hierarchical Communities Qing Song, Xiaofan Wang Department of Automation, Shanghai Jiao Tong University,
How to Cut Pseudoparabolas into Segments Seminar on Geometric Incidences By: Almog Freizeit.
CSE 589 Part VI. Reading Skiena, Sections 5.5 and 6.8 CLR, chapter 37.
PaGrid: A Mesh Partitioner for Computational Grids Virendra C. Bhavsar Professor and Dean Faculty of Computer Science UNB, Fredericton This.
1 Partitioning. 2 Decomposition of a complex system into smaller subsystems  Done hierarchically  Partitioning done until each subsystem has manageable.
CS 484 Load Balancing. Goal: All processors working all the time Efficiency of 1 Distribute the load (work) to meet the goal Two types of load balancing.
Adaptive Mesh Applications Sathish Vadhiyar Sources: - Schloegel, Karypis, Kumar. Multilevel Diffusion Schemes for Repartitioning of Adaptive Meshes. JPDC.
Partitioning using Mesh Adjacencies  Graph-based dynamic balancing Parallel construction and balancing of standard partition graph with small cuts takes.
Domain decomposition in parallel computing Ashok Srinivasan Florida State University.
Data Structures and Algorithms in Parallel Computing Lecture 7.
ICS 252 Introduction to Computer Design
CprE566 / Fall 06 / Prepared by Chris ChuPartitioning1 CprE566 Partitioning.
Large Scale Parallel Graph Coloring 1. Presentation Overview Problem Description Basic Algorithm Parallel Strategy –Work Spawning –Graph Partition Results.
Global Clustering-Based Performance-Driven Circuit Partitioning Jason Cong University of California Los Angeles Chang Wu Aplus Design.
Predictive Load Balancing Using Mesh Adjacencies for Mesh Adaptation  Cameron Smith, Onkar Sahni, Mark S. Shephard  Scientific Computation Research Center.
Hypergraph Partitioning With Fixed Vertices Andrew E. Caldwell, Andrew B. Kahng and Igor L. Markov UCLA Computer Science Department
Multilevel Partitioning
Graphs Definition: a graph is an abstract representation of a set of objects where some pairs of the objects are connected by links. The interconnected.
University of Texas at Arlington Scheduling and Load Balancing on the NASA Information Power Grid Sajal K. Das, Shailendra Kumar, Manish Arora Department.
3/21/ VLSI Physical Design Automation Prof. David Pan Office: ACES Lecture 4. Circuit Partitioning (II)
Mesh Generation, Refinement and Partitioning Algorithms Xin Sui The University of Texas at Austin.
CS 140: Sparse Matrix-Vector Multiplication and Graph Partitioning
Partitioning Jong-Wha Chong Wireless Location and SOC Lab. Hanyang University.
High Performance Computing Seminar II Parallel mesh partitioning with ParMETIS Parallel iterative solvers with Hypre M.Sc. Caroline Mendonça Costa.
High Performance Computing Seminar
Parallel Hypergraph Partitioning for Scientific Computing
Parallel Programming By J. H. Wang May 2, 2017.
A Continuous Optimization Approach to the Minimum Bisection Problem
Plan Introduction to multilevel heuristics Rich partitioning problems
Using Multilevel Force-Directed Algorithm to Draw Large Clustered Graph 研究生: 何明彥 指導老師:顏嗣鈞 教授 2018/12/4 NTUEE.
Multi-level Coarsening Algorithm
Design Hierarchy Guided Multilevel Circuit Partitioning
Integrating Efficient Partitioning Techniques for Graph Oriented Applications My dissertation work represents a study of load balancing and data locality.
A Fundamental Bi-partition Algorithm of Kernighan-Lin
A Parallelization of State-of-the-Art Graph Bisection Algorithms
Rusakov A. S. (IPPM RAS), Sheblaev M.
Fast Min-Register Retiming Through Binary Max-Flow
Presentation transcript:

Multilevel Hypergraph Partitioning G. Karypis, R. Aggarwal, V. Kumar, and S. Shekhar Computer Science Department, U of MN Applications in VLSI Domain

Overview Problem definition Earlier work on partitioning Multilevel graph partitioning algorithms Multilevel hypergraph partitioning algorithms Experimental results Conclusion

Problem Definition Given a hypergraph G = (V, E) we want to find a balanced bisection such that the number of hyperedges that are cut is minimized. Applications 4 VLSI (packaging, synthesis, optimization, …) 4 Numerical computations 4 Transportation 4 Data-mining

Overview of Partitioning Algorithms Effective partitioning algorithms must utilize both global as well as local information. Global View Knowing where to bisect Local ViewThe ability to fine-tune a bisection A number of partitioning algorithms have been developed, utilizing global and local information to different degrees. Single-level Two-level Multi-level

Multilevel Partitioning Schemes l Originally developed for graphs (edges connecting 2 nodes) as opposed to netlists or hypergraphs in which edges or hyperedges connect >= 2 nodes) l Produce high-quality partitionings. Incorporate both global and local information. Outperforms everything else! l They are extremely fast. Eg: 1M-node graph takes 35s. l They can be easily parallelized. Eg: 1M-node graph takes 0.8s on 64 processors. Coarsening Uncoarsening + Partition Refinement Initial Partitioning

Ingredients of Multilevel Partitioning CoarseningInitial PartitioningRefinement u Successive coarse graphs must make it easier to find a good partition. u Uniform vertex weights (node/vertex “sizes” should be as uniform as possible). u Exposed edge-weight must decrease rapidly. u The `how to coarsen’ computation must be fast. u The size of successive coarse graphs must decrease relatively fast u less time spent in coarsening, less memory.

Ingredients of Multilevel Partitioning CoarseningInitial PartitioningRefinement u This is the easiest of the three phases. u Everything reasonable works fine. Random+FM, spectral, region growing, etc. u It requires very little time Operates on small graphs (~100 vertices).

Ingredients of Multilevel Partitioning CoarseningInitial PartitioningRefinement u Needs a local partitioning refinement algorithm. u Any vertex-swapping algorithm can be used KL, FM, etc. u If coarsening is done correctly, simple refinement algorithms work extremely well and this phase requires very little time.

Metis: Multilevel Graph Partitioning u Coarsening u Maximal independent set of edges (matching). u Preference to high weight edges: heavy-edge. u Effective in reducing the exposed edge-weight! u Initial Partitioning u A region-growing followed by FM u Refinement u A simplified version of FM Only up to 4 passes, Early exit u Very fast refinement. Metis is an extremely fast, robust, high-quality graph partitioning algorithm

Going from Graphs to Hypergraphs Hypergraph partitioning is significantly more complicated than graph partitioning. Just look at the various refinement algorithms used in hypergraphs. n Graphs: KL/FM n Hypergraphs: KL/FM, LA, PROP, CLIP, etc. Can we find proper coarsening schemes that will let us use simple and fast refinement schemes and get good and robust performance?

Hypergraph Coarsening Schemes Edge-based coarsening schemes n Pairs of connected vertices are collapsed together, using the heavy-edge heuristic. 4 Easy and fast to compute. 8 Does not dramatically decrease the exposed hyperedge weight. u Cannot easily remove moderate-size hyperedges. 8 Requires a lot of refinement in order to obtain good partitionings u Requires sophisticated refinement schemes 8 Can lead to good partitionings but very slow!

Hypergraph Coarsening Schemes Hyperedge-based coarsening schemes n Collapses together all the vertices of an entire hyperedge. n Preference is given to the heavier hyperedges. 4 Easy and fast to compute. 4 It dramatically decreases the exposed hyperedge weight. 4 Leads to very good initial partitionings. u 3600 as opposed to 6200 for golem3! 4 Requires very little refinement time. 4 High-quality partitionings can be obtained with simple refinement schemes.

hMetis: Multilevel Hypergraph Partitioning Algorithm Uses hyperedge-based coarsening. Uses a simplified version of FM for refinement n Limits the number of passes, Early-exit Employs some new multilevel refinement techniques to further improve the quality. hMetis is an extremely fast, robust, high- quality hypergraph partitioning algorithm.

Experimental Setup We used the ACM/SIGDA circuit partitioning benchmark. Experiments were performed on a MIPS Results represent a balance condition. Best out of 20 runs using EE-FM and multilevel refinement.

Bisection Quality

Bisection Runtime

Conclusions The multilevel paradigm with the right coarsening and refinement scheme works extremely well for hypergraphs. The quality of the partitionings can be further improved by running the algorithm multiple times. hMetis will be made available in the public domain by the end of June. URL: