Abstract. Graph coloring has been employed since the 1980s to efficiently compute sparse Jacobian and Hessian matrices using either finite differences or automatic differentiation. Several coloring problems occur in this context, depending on whether the matrix is a Jacobian or a Hessian, and on the specifics of the computational techniques employed. We consider eight variant vertex coloring problems here. This article begins with a gentle introduction to the problem of computing a sparse Jacobian, followed by an overview of the historical development of the research area. Then we present a unifying framework for the graph models of the variant matrix estimation problems. The framework is based upon the viewpoint that a partition of a matrix into structurally orthogonal groups of columns corresponds to distance-2 coloring an appropriate graph representation. The unified framework helps integrate earlier work and leads to fresh insights; enables the design of more efficient algorithms for many problems; leads to new algorithms for others; and eases the task of building graph models for new problems. We report computational results on two of the coloring problems to support our claims. Most of the methods for these problems treat a column or a row of a matrix as an atomic entity, and partition the columns or rows (or both). A brief review of methods that do not fit these criteria is provided. We also discuss results in discrete mathematics and theoretical computer science that intersect with the topics considered here.

SUMMARYFinding a good graph coloring quickly is often a crucial phase in the development of efficient, parallel algorithms for many scientific and engineering applications. In this paper we consider the problem of solving the graph coloring problem itself in parallel. We present a simple and fast parallel graph coloring heuristic that is well suited for shared memory programming and yields an almost linear speedup on the PRAM model. We also present a second heuristic that improves on the number of colors used. The heuristics have been implemented using OpenMP. Experiments conducted on an SGI Cray Origin 2000 supercomputer using very large graphs from finite element methods and eigenvalue computations validate the theoretical run-time analysis.

Abstract-DBSCAN is a well-known density based clustering algorithm capable of discovering arbitrary shaped clusters and eliminating noise data. However, parallelization of DBSCAN is challenging as it exhibits an inherent sequential data access order. Moreover, existing parallel implementations adopt a master-slave strategy which can easily cause an unbalanced workload and hence result in low parallel efficiency.We present a new parallel DBSCAN algorithm (PDSDBSCAN) using graph algorithmic concepts. More specifically, we employ the disjoint-set data structure to break the access sequentiality of DBSCAN. In addition, we use a tree-based bottom-up approach to construct the clusters. This yields a better-balanced workload distribution. We implement the algorithm both for shared and for distributed memory.Using data sets containing up to several hundred million high-dimensional points, we show that PDSDBSCAN significantly outperforms the master-slave approach, achieving speedups up to 25.97 using 40 cores on shared memory architecture, and speedups up to 5,765 using 8,192 cores on distributed memory architecture.

No abstract

We present new multi-core algorithms for computing spanning forests and connected components of large sparse graphs. The algorithms are based on the use of the disjoint-set data structure. When compared with the previous best algorithms for these problems our algorithms are appealing for several reasons: Extensive experiments using up to 40 threads on several different types of graphs show that they scale better. Also, the new algorithms do not make use of any hardware specific routines, and thus are highly portable. Finally, the algorithms are quite simple and easy to implement.

It is a well-established result that improved pivoting in linear solvers can be achieved by computing a bipartite matching between matrix entries and positions on the main diagonal. With the availability of increasingly faster linear solvers, the speed of bipartite matching computations must keep up to avoid slowing down the main computation. Fast algorithms for bipartite matching, which are usually initialized with simple heuristics, have been known for a long time. However, the performance of these algorithms is largely dependent on the quality of the heuristic. We compare combinations of several known heuristics and exact algorithms to find fast combined methods, using real-world matrices as well as randomly generated instances. In addition, we present a new heuristic aimed at obtaining high-quality matchings and compare its impact on bipartite matching algorithms with that of other heuristics. The experiments suggest that its performance compares favorably to the best-known heuristics, and that it is especially suited for application in linear solvers.

No abstract

Matrix partitioning problems that arise in the efficient estimation of sparse Jacobians and Hessians can be modeled using variants of graph coloring problems. In a previous work [7], we argue that distance-2 and distance-3 2 graph coloring are robust and flexible formulations of the respective matrix estimation problems. The problem size in large-scale optimization contexts makes the matrix estimation phase an expensive part of the entire computation both in terms of execution time and memory space. Hence, there is a need for both shared-and distributed-memory parallel algorithms for the stated graph coloring problems. In the current work, we present the first practical shared address space parallel algorithms for these problems. The main idea in our algorithms is to randomly partition the vertex set equally among the available processors, let each processor speculatively color its vertices using information about already colored vertices, detect eventual conflicts in parallel, and finally re-color conflicting vertices sequentially. Randomization is also used in the coloring phases to further reduce conflicts. Our PRAM-analysis shows that the algorithms should give almost linear speedup for sparse graphs that are large relative to the number of processors. Experimental results from our OpenMP implementations on a Cray Origin2000 using various large graphs show that the algorithms indeed yield reasonable speedup for modest numbers of processors.