Masoumeh Moeini - School of Electrical Engineering Iran University of Science and Technology Tehran, Iran
Hadi Shahriar Shahhoseini - School of Electrical Engineering Iran University of Science and Technology Tehran, Iran

String matching algorithms play an important role in many aspects. In this paper, a new method is proposed for parallel execution of Rabin-Karp string matching algorithm. In the proposed method, all patterns are divided into different groups. This categorization has been used to prevent redundant comparisons and accelerate the matching process. This procedure takes two advantages: a reduction in the number of comparisons of hash values and a decline in the number of pattern reading from global memory. It is recommended to implement the algorithm in various cases on NVIDIA GPUs using CUDA Platform to demonstrate the efficiency of the proposed algorithm. Experimental results depict that the execution time of the algorithm has decreased significantly. In the best case, the proposed method is approximately 27 times faster than the series mode