List ranking and list scan are two primitive operations used in many parallel algorithms that use list, trees, and graph data structures. But vectorizing and parallelizing list ranking is a challenge because it is highly communication intensive and dynamic. In addition, the serial algorithm is very simple and has very small constants. In order to compete, a parallel algorithm must also be simple and have small constants. A parallel algorithm due to Wyllie is such an algorithm, but it is not work efficient—its performance degrades for longer and longer linked lists. In contrast, work efficient PRAM algorithms developed to date have very large constants. It does not achieve O(log n) running time, but we contend that work efficiency and small constants is more important, given that vector and multiprocessor machines are used for problems that are much larger than the number of processors and, therefore, the O(log n) running time, but we contend that work efficiency and small constants is more important, given that vector and multiprocessor machines are used for problems that are much larger than the number of processors and, therefore, the O(log n) time is never achieved in practice. In particular, to the best of our knowledge, our implementation of list ranking and list scan on the CRAY C-90 is the fastest implementation to date. In addition, it is the first implementation of which we are aware that outperforms fast workstations. The success of our algorithm is due to its relatively large grain size and simplicity of the inner loops, and the success of the implementation is due to pipelining reads and writes through vectorization to hide latency, minimizing load balancing by deriving equations for predicting and optimizing performance, and avoiding conditional tests except when load balancing.

List ranking and list scan on the Cray C-90

The minimum spanning tree problem with an added constraint that no node in the spanning tree has the degree more than a specified integer, d, is known as the minimum-weight degree-constrained spanning tree (d-MST) problem. Such a constraint arises, for example, in VLSI routing trees, in backplane wiring, or in minimizing single-point failures for communication networks. The d-MST problem is NP-complete. Here, we develop four heuristics for approximate solutions to the problem and implement them on a massivelyparallel SIMD machine, MasPar MP-1. An extensive empirical study shows that for random graphs on up to 5000 nodes (about 12.5 million edges), the heuristics produce solutions close to the optimal in less than 10 seconds. The heuristics were also tested on a number of TSP benchmark problems to compute spanning trees with a degree bound d = 3.

Minimum-weight degree-constrained spanning tree problem: heuristics and implementation on an SIMD parallel machine

We describe the implementation of virtual processing for several combinatorial algorithms using the MPL language on a 16,384-processor MasPar MP-1. In coding these routines, we tried different underlying (deterministic and randomized) algorithms. We present the performance data for our different implementations. We also present general code rewriting rules for converting a code that uses no virtual processor into a code with virtual processing.

V. Ramaehandran

Papers

Efficient implementation of virtual processing for some combinatorial algorithms on the MasPar MP-1