Cascading Divide-and-Conquer: A Technique for Designing Parallel Algorithms

We present techniques which result in improved parallel algorithms for a number of problems whose efficient sequential algorithms use the plane-sweeping paradigm. The problems for which we give improved algorithms include intersection detection, trapezoidal decomposition, triangulation, and planar point location. Our technique can be used to improve on the previous time bound while keeping the space and processor bounds the same, or improve on the previous space bound while keeping the time and processor bounds the same. We also give efficient parallel algorithms for visibility from a point, 3-dimensional maxima, multiple range-counting, and rectilinear segment intersection counting. We never use the AKS sorting network in any of our algorithms.

Efficient parallel algorithms are presented, on the CREW PRAM model, for generating a succinct encoding of all pairs shortest path information in a directed planar graphG with real-valued edge costs but no negative cycles. We assume that a planar embedding ofG is given, together with a set ofq faces that cover all the vertices. Then our algorithm runs inO(log2 n) time and employsO(nq+M(q)) processors (whereM(t) is the number of processors required to multiply twot×t matrices inO(logt) time). Let us note here that wheneverq<n then our processor bound is better than the best previous one (M(n)).O(log2 n) time,n-processor algorithms are presented for various subproblems, including that of generating all pairs shortest path information in a directedouterplanar graph. Our work is based on the fundamental hammock-decomposition technique of G. Frederickson. We achieve this decomposition inO(logn log*n) parallel time by usingO(n) processors. The hammock-decomposition seems to be a fundamental operation that may help in improving efficiency of many parallel (and sequential) graph algorithms.

PARALLEL AND LARGE SCALE MATRIX COMPUTATIONS IN CONTROL: SOME IDEAS

Further applications of random sampling techniques which have been used for deriving efficient parallel algorithms are presented by J. H. Reif and S. Sen [Proc. 16th International Conference on Parallel Processing, 1987]. This paper presents an optimal parallel randomized algorithm for computing intersection of half spaces in three dimensions. Because of well-known reductions, these methods also yield equally efficient algorithms for fundamental problems like the convex hull in three dimensions, Voronoi diagram of point sites on a plane, and Euclidean minimal spanning tree. The algorithms run in time $T = O(\log n)$ for worst-case inputs and use $P = O(n)$ processors in a CREW PRAM model where n is the input size. They are randomized in the sense that they use a total of only polylogarithmic number of random bits and terminate in the claimed time bound with probability $1 - n^{ - \alpha } $ for any fixed $\alpha > 0$. They are also optimal in $P\cdot T$ product since the sequential time bound for all these problems is $\Omega (n\log n)$. The best known deterministic parallel algorithms for two-dimensional Voronoi-diagram and three-dimensional convex hull run in $O(\log ^2 n)$ and $O(\log ^2 n\log ^ * n)$ time, respectively, while using $O(n/\log n)$ and $O(n)$ processors, respectively.

In this paper we study the question: How useful is randomization in speeding up Exclusive Write PRAM computations? Our results give further evidence that randomization is of limited use in these types of computations. First we examine a compaction problem on both the CREW and EREW PRAM models, and we present randomized lower bounds which match the best deterministic lower bounds known. (For the CREW PRAM model, the lower bound is asymptotically optimal.) These are the first non-trivial randomized lower bounds known for the compaction problem on these models. We show that our lower bounds also apply to the problem of approximate compaction. Next we examine the problem of computing boolean functions on the CREW PRAM model, and we present a randomized lower bound, which improves on the previous best randomized lower bound for many boolean functions, including the OR function. (The previous lower bounds for these functions were asymptotically optimal, but we improve the constant multiplicative factor.) We also give an alternate proof for the randomized lower bound on PARITY, which was already optimal to within a constant additive factor. Lastly, we give a randomized lower bound for integer merging on an EREW PRAM which matches the best deterministic lower bound known. In all our proofs, we use the Random Adversary method, which has previously only been used for proving lower bounds on models with Concurrent Write capabilities. Thus this paper also serves to illustrate the power and generality of this method for proving parallel randomized lower bounds.

Operator Splitting and Domain Decomposition for Multiclusters

High-performance computing for climate models has always been an interesting research area. It is valuable to nest a regional climate model within a global climate model, but large-scale simulation of the nesting or coupling severely challenges to the development of efficient parallel algorithms that fit well into multi-core clusters. This paper first presents research on the coupling of the Institute of Atmospheric Physics of Chinese Academy of Sciences Atmospheric General Circulation Model version 4.0 and the Weather Research and Forecasting model, then proposes an efficient parallel algorithm of the coupling. The algorithm includes initialization of input data, decomposition of computing grid and processes, parallel computing of component models, and data exchange by a coupler. By calling some subroutines of the Model Coupling Toolkit, the parallelization of the proposed algorithm is implemented. Experiments show that the parallel algorithm is very effective and scalable. The parallel efficiency of the algorithm on 1,024 CPU cores can reach up to 70%. Moreover, its parallel efficiency with respect to weak scalability is 72.56% on a multi-core cluster.

Over the years many interesting and efficient parallel algorithms have been developed to solve a wide variety of problems, but not much attention has been paid to studying the inherent parallelism in sequential algorithms---i.e., understanding the depth of their dependence structure, and how shallow dependence structures might beused to develop efficient parallel implementations. In this talk I will describe recent work on analyzing the dependence depth of iterative sequential algorithms---ones that loop over a collection of elements. Many of these algorithms have deep dependence chains in the worst case, but shallow chains (polylog w.h.p.) if the elements are randomly ordered. Examples include many fundamental algorithms: the Knuth shuffle for random permutations, sorting by insertion into a binary search tree, greedy maximal independent set (MIS), greedy maximal matching, greedy graph-coloring, counting cycles in a permutation, incremental k-dimensional linear programming, and incremental 2d Delaunay triangulation. An advantage of the approach is that it can lead to very simple and efficient parallel algorithms. Our MIS algorithm, for example can be coded in a dozen or so lines, and is significantly faster than Luby's algorithm on modern multicore machines. Also the approach encourages snapping the view that sequential and parallel algorithms are distinct, and instead thinking of algorithms, in general, as collections of instructions with dependences among them.

The all-pairs shortest path problem is a classic problem to study characteristics of the given graphs. Though many efficient all-pairs shortest path algorithms have been published, it is still a very expensive computing task, especially with large graph datasets. In this paper, we propose an efficient parallel all-pairs shortest path algorithm based on Peng et al.'s fast sequential algorithm on shared-memory parallel environments to achieve faster and more efficient calculation for large-scale real-world networks. Peng et al.'s algorithm needs to sort vertices with respect to their degrees. However, it turns out the original algorithm uses less efficient sorting method, which is a significant portion of parallel overhead. Therefore, we also propose an efficient parallel method to sort data within a fixed range, in order to minimize the parallel overhead in our parallel algorithm. The optimized efficient sorting method can be used for general sorting purposes. Our experimental analysis shows that our proposed parallel algorithm achieves very high parallel speedup, even hyper-linear speedup, with real-world test datasets on two different shared-memory multi-core systems.

An efficient parallel algorithm is presented for computing selected components of $A^{-1}$ where A is a structured symmetric sparse matrix. Calculations of this type are useful for several applications, including electronic structure analysis of materials in which the diagonal elements of the Green's functions are needed. The algorithm proposed here is a direct method based on a block $LDL^T$ factorization. The selected elements of $A^{-1}$ we compute lie in the nonzero positions of $L+L^T$. We use the elimination tree associated with the block $LDL^T$ factorization to organize the parallel algorithm, and reduce the synchronization overhead by passing the data level by level along this tree using the technique of local buffers and relative indices. We demonstrate the efficiency of our parallel implementation by applying it to a discretized two dimensional Hamiltonian matrix. We analyze the performance of the parallel algorithm by examining its load balance and communication overhead, and show that our parallel implementation exhibits an excellent weak scaling on a large-scale high performance distributed-memory parallel machine.

A node ranking of a graph G = (V , E ) is a proper node coloring C : V ****** such that any path in G with end nodes x , y fulfilling C (x ) = C (y ) contains an internal node z with C (z ) > C (x ). In the on-line version of the node ranking problem, the nodes v 1 , v 2 ,..., v n are coming one by one in an arbitrary order; and only the edges of the induced subgraph G [{v 1 , v 2 ,..., v i }] are known when the color for the node v i be chosen. And the assigned color can not be changed later. In this paper, we present a parallel algorithm to find an on-line node ranking for general tree. Our parallel algorithm needs O (n log2 n ) time with using O (n 3 / log2 n ) processors on CREW PRAM model.

Efficient parallel algorithms and software for compressed octrees with applications to hierarchical methods

Efficient parallel and fast convergence chaotic Jaya algorithms

An efficient parallel algorithm for testing whether a graph G is k-vertex connected is presented. The algorithm runs in $O(k^2 \log n)$ time and uses $(n + k^2 )kC(n,m)$ processors on a CROW PRAM, where n and m are the number of vertices and edges of G, and $C(n,m)$ is the number of processors required to compute the connected components of G in logarithmic time. For fixed k, the algorithm runs in logarithmic time and uses $nC(n,m)$ processors. To develop our algorithm, an efficient parallel algorithm is designed for the following disjoint s-t paths problem. Given a graph G, and two specified vertices s and t, find k vertex disjoint paths between s and t, if they exist. If no such paths exist, find a set of at most $k-1$ vertices whose removal disconnects s and t. The parallel algorithm for this problem runs in $O(k^2 \log n)$ time and uses $kC(n,m)$ processors. The way to modify the algorithm to find k-edge disjoint paths, if they exist, is shown. This yields an efficient parallel algorithm for testing w...

The efficient parallel algorithms proposed for many fundamental problems, such as list ranking, computing preorder numberings on trees and integer sorting, are very sensitive to processor failures. The requirements of efficiency (commonly formalized using {\em Parallel-time x Processors} as a cost measure) has led to the design of highly tuned PRAM algorithms which, given the additional constraint of simple processor failures, unfortunately become inefficient or even incorrect. We propose a new notion of {\em robustness} that combines efficiency with fault tolerance. For the common case of fail-stop errors, we develop a general and easy-to-implement technique to make robust many efficient parallel algorithms, e.g., algorithms for all the problems listed above. More specifically,{\em for any dynamic pattern of fail-stop errors on} a CRCW PRAM {\em with at least one surviving processor}, our method increases the original algorithm''s cost by at most a {\em polylog} multiplicative factor. We also show that at least a {\em log/loglog} multiplicative overhead will be incurred for certain patterns of failures by any algorithm that implements robust solutions to the problems listed.