Pseudo associative linking: a high speed searching algorithm for parallel processors
The task of correlating a large number of radar reports with a large number of stored radar tracks has long been identified as a natural application of single instruction multiple data computers. As practical/deployable parallel processors are often limited in power and number of processing elements, searching techniques have been sought that would provide faster execution times than that obtained by brute force searches in an SIMD machine. Specifically the authors sought some means of utilizing the power of the parallel processor in the creation and searching of a data structure that would give a species of associativity such as enjoyed by a machine with a far larger number of processing elements (pes). A data structure and set of searching algorithms have been developed which they call pseudo associative linking (PAL). With this technique one can, in effect, multiply the power of the SIMD machine on itself. The developed algorithms have been successfully used in the implementation of the Litton Advanced Tracking System (ATS) developed for Rome Air Development Center which uses the 16 element tracking array processor (TAP) as the parallel processor. The Litton ATS has already demonstrated a track processing capability of over 10000 tracks. 5 references.
- OSTI ID:
- 5180049
- Report Number(s):
- CONF-8308190-
- Resource Relation:
- Conference: 12. IEEE international conference on parallel processing, Bellaire, MI, USA, 21 Aug 1983
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallelized Kalman-Filter-Based Reconstruction of Particle Tracks on Many-Core Processors and GPUs
Scaling molecular dynamics beyond 100,000 processor cores for large-scale biophysical simulations