A Matrix Partitioning Interface to PaToH in MATLAB

Uçar, Bora; Çatalyürek, Ümit V.; Aykanat, Cevdet

doi:10.1016/j.parco.2009.12.008

Cited by 12 publications

(12 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We ran our tests using PaToH Matlab Matrix-Partitioning Interface [11], [12] on a dual quad-core 2.26 GHz Intel Xeon desktop with 24 GB of memory using Matlab v7.8 (R2009a).…”

Section: Resultsmentioning

confidence: 99%

On the Scalability of Hypergraph Models for Sparse Matrix Partitioning

Uçar

Çatalyürek

2010

2010 18th Euromicro Conference on Parallel, Distributed and Network-Based Processing

View full text Add to dashboard Cite

To cite this version:Bora Uçar, Umit Catalyurek. On the scalability of hypergraph models for sparse matrix partitioning. M. Danelutto, J. Bourgeois, and T. Gross. Abstract-We investigate the scalability of the hypergraphbased sparse matrix partitioning methods with respect to the increasing sizes of matrices and number of nonzeros. We propose a method to rowwise partition the matrices that correspond to the discretization of two-dimensional domains with the five-point stencil. The proposed method obtains perfect load balance and achieves very good total communication volume. We investigate the behaviour of the hypergraph-based rowwise partitioning method with respect to the proposed method, in an attempt to understand how scalable the former method is. In another set of experiments, we work on general sparse matrices under different scenarios to understand the scalability of various hypergraphbased one-and two-dimensional matrix partitioning methods.

show abstract

“…We ran our tests using PaToH Matlab Matrix-Partitioning Interface [11], [12] on a dual quad-core 2.26 GHz Intel Xeon desktop with 24 GB of memory using Matlab v7.8 (R2009a).…”

Section: Resultsmentioning

confidence: 99%

On the Scalability of Hypergraph Models for Sparse Matrix Partitioning

Uçar

Çatalyürek

2010

2010 18th Euromicro Conference on Parallel, Distributed and Network-Based Processing

View full text Add to dashboard Cite

show abstract

“…We developed a matrix--partitioning interface to PaToH in Matlab [15] for partitioning sparse matrices. This allows Matlab users to partition their sparse matrices using 1D and 2D partitioning techniques we developed [16].…”

Section: Matrix Partitioningmentioning

confidence: 99%

SciDAC Institute: Combinatorial Scientific Computing and Petascale Simulations (CSCAPES) (Final Report)

Çatalyürek¹

2014

View full text Add to dashboard Cite

“…For an optimal solution of an instance of the online partitioning problem, we utilize the -way hypergraph partitioning algorithm [4]. Given an instance of the problem, we can construct a hypergraph such that the partitioning of the nodes using the connectivity criterion yields the optimal solution for that instance.…”

Section: Optimal Solutionmentioning

confidence: 99%

“…Unfortunately, this is infeasible in practical scenarios as hypergraph partitioning is NP-hard [4]. Even with heuristic hypergraph partitioning algorithms, the solution comes with considerable overhead for two reasons.…”

Section: Optimal Solutionmentioning

confidence: 99%

Online horizontal partitioning of heterogeneous data

Herrmann

Voigt

Lehner

2014

It – Information Technology

View full text Add to dashboard Cite

In an increasing number of use cases, databases face the challenge of managing heterogeneous data. Heterogeneous data is characterized by a quickly evolving variety of entities without a common set of attributes. These entities do not show enough regularity to be captured in a traditional database schema. A common solution is to centralize the diverse entities in a universal table. Usually, this leads to a very sparse table. Although today's techniques allow efficient storage of sparse universal tables, query efficiency is still a problem. Queries that address only a subset of attributes have to read the whole universal table including many irrelevant entities. A solution is to use a partitioning of the table, which allows pruning partitions of irrelevant entities before they are touched. Creating and maintaining such a partitioning manually is very laborious or even infeasible, due to the enormous complexity. Thus an autonomous solution is desirable. In this article, we define the Online Partitioning Problem for heterogeneous data. We sketch how an optimal solution for this problem can be determined based on hypergraph partitioning. Although it leads to the optimal partitioning, the hypergraph approach is inappropriate for an implementation in a database system. We present Cinderella, an autonomous online algorithm for horizontal partitioning of heterogeneous entities in universal tables. Cinderella is designed to keep its overhead low by operating online; it incrementally assigns entities to partition while they are touched anyway during modifications. This enables a reasonable physical database design at runtime instead of static modeling.

show abstract

A Matrix Partitioning Interface to PaToH in MATLAB

Cited by 12 publications

References 27 publications

On the Scalability of Hypergraph Models for Sparse Matrix Partitioning

On the Scalability of Hypergraph Models for Sparse Matrix Partitioning

SciDAC Institute: Combinatorial Scientific Computing and Petascale Simulations (CSCAPES) (Final Report)

Online horizontal partitioning of heterogeneous data

Contact Info

Product

Resources

About