The optimal running time from computational models by comparing two parallel searching procedures

This paper proposes a parallel implementation of search algorithms on SIMD computers to be efficiently executed. In these parallel searching, the desired data is queried in the random sequence. Objectives of this paper are two-fold; the first demonstrates the SM SEARCH (S, x, k) algorithm in shared memory SIMD computers and MESH SEARCH (S, x, answer) in mesh-connected SIMD computers. The second is to analyze these two procedures in CREW and EREW computational models respectively and compares the running time of these models to get the optimal running time. The querying on data of computer universities will be played in a key role to provide the implementation of two parallel searching methods. This experimental results prove that parallel implementation of MESH SEARCH on a mesh-connected SIMD computer achieves significant speedup and performance over SM SIMD computer.