Platform Independent PFAC Implementations using OpenCL on Heterogeneous Parallel Computing

The basic and standard multiple patterns string matching algorithm is Aho-Corasick invented by Alfred V. Aho and Margaret J. Corasick. The algorithm of Aho-Corasick can match multiple patterns simultaneously and affirmed deterministic performance under all circumstances. Various real world applications were provided by this algorithm like computational biology, intrusion detection systems, multimedia, search engine and text mining. Performances parallelization of Aho-Corasick is crucial in order to improve performance of these applications and meet with real time environments. Parallelization of Aho-Corasick can provide drastic performance under various circumstances. The parallel version of Aho-Corasick is PFAC (Parallel Failure Less Aho-Corasick). PFAC algorithm provides high degrees of parallelization and various improvements in Aho-Corasick algorithm. We are going to implement PFAC. Along with PFAC various memory access techniques will be used to implement the algorithm. Also to bring to notice that previously PFAC library is built in CUDA. These libraries are platform dependent and run on only NVIDIA GPUs. General-purpose computing on graphics processing units (GPGPU) is the utilization of a graphics processing unit (GPU), which typically handles computation only for computer graphics, to perform computation in applications traditionally handled by the central processing unit (CPU). In this paper we have various versions of PFAC using OpenCL which is platform independent and able to run on any GPU. Here we will also discuss the performance details of various PFAC versions