Parallel Frame-Based Signal Filtering on Shared-Memory Multi-Core CPU Using OpenMP

Abstract

Conference Title: 2025 IEEE XVII International Scientific and Technical Conference on Actual Problems of Electronic Instrument Engineering (APEIE)

Conference Start Date: 2025 Nov. 14

Conference End Date: 2025 Nov. 16

Conference Location: Novosibirsk, Russian Federation

To achieve high speed and efficiency in digital data processing, it is important to choose the right computing resources. It is known that specialized digital signal processors and graphics processors for processing data such as signals exist, but these computing resources are not available in all studies. This study examines the possibility of maximizing the use of available resources in cases where specialized hardware resources for signal processing are insufficient. In this case, it is advisable to use a modern multi-core processor designed for general-purpose tasks. However, since the processors are adapted for sequential processing, parallel processing of signal filtering processes is performed in this processor. Since the processor can provide a sufficient level of parallel processing based on the L1, L2, and L3 cache memory systems, the Open Multi-Processing programming model, with its capabilities, can significantly accelerate signal filtering processes performed by parallel processing tools in processors with shared memory. In this case, the signal data is divided into frames of 8, 16, 32,... 2048, and each frame is evenly distributed among the processor cores using Open Multi-Processing, and parallel calculations are performed simultaneously. When implementing parallel filtering of the moving average and root mean square methods for signal filtering on a 4-core (8-thread) processor, the acceleration was ~1.2-2.5 times for small frames, ~3.5-3.8 times for medium frames, and ~4.4 times for large frames.