Performance Evaluation of Regular Expression Matching Engines Across Different Computer Architectures

Regular expressions are sequences of characters that define search patterns, commonly used in pattern matching with strings. Regular expression matching plays an important role in a variety of applications, such as bioinformatics, network inspection, etc. However, it is a challenging problem because pattern matching is a computationally intensive operation especially when dealing with large data and complex regular expression rulesets. In this work we investigate for a fair comparison of regular expression matching engines across parallel architectures, including multi-core central processing units (CPUs), SIMD-based graphics processing units (GPUs), many-core Intel’s XeonPhi accelerators, field-programmable gate arrays (FPGAs), and Automata Processor (AP). The evaluation involves analyzing their performances with different types of regular expressions and exploring the design spaces of these architectures: “complexity” of regular expressions, the number of regular expressions, and multiple packets processing capability. The work is also extended to other applications that are not natural fits for regular expression. Keywords— Automata Processor (AP); central processing unit (CPU); graphics processing unit (GPU); field-programmable gate array (FPGA); finite automata; parallel architectures; regular expression matching; Intel’s XeonPhi accelerator