Parallel tree search on a SIMD machine

Presents an implementation of a depth-first heuristic tree search on the single-instruction, multiple-data (SIMD) Connection Machine. The algorithm is based on Iterative-Deepening-A*. Until recently, only highly regular, data-parallel computations have been performed on SIMD machines. Searching an irregular tree represents a new application of SIMD machines. The main technical challenge is load balancing, and the authors explore three different techniques in combination. They also present a simple and general method for dynamically determining when to stop working and start load balancing. They achieve an efficiency of 67%, for a speedup of 5400, on 8 K processors, and an efficiency of 57%, for a speedup of 9300, on 16 K processors.<
>