Benchmarking Complex Multiplication in Signal Processing Using Intel OneAPI: A Comparative Study on CPU, GPU and FPGA Performance
T.S. Balaji, B. Bhuvanesan, C. Vishnu, G. Revanth, Anas Akhter · 2024
The efficiency of complex multiplication in signal processing is investigated in this study on three hardware platforms: CPU, GPU, and FPGA. A unified codebase was created to assess and contrast the efficiency and speed of execution on each platform by utilizing Intel oneAPI. The analysis highlights the disparities in parallel processing capabilities between conventional CPUs and the specialized architectures of GPUsand FPGAs, with a particular emphasis on their performance in intricate multiplication tasks. The findings offer practical insights into the selection of the most suitable hardware for real-time signal processing, elucidating the strengths and limitations ofeach platform.