Citation :

Muthukumaran Vaithianathan, Mahesh Patil, Shunyee Frank Ng, Shiv Udkar, 2023. "Comparative Study of FPGA and GPU for High-Performance Computing and AI" ESP International Journal of Advancements in Computational Technology (ESP-IJACT)  Volume 1, Issue 1: 37-46.

Abstract :

The complexity of computing problems has made it possible for researchers to seek different computational environments to achieve optimum performance in high-performance computing and artificial intelligence. In this context, Field Programmable Gate Arrays (FPGAs) and Graphics Processing Units (GPUs) are now seen as key technologies as each has its strengths. This paper tries to compare the FPGAs and GPUs focusing on the different aspects like performance, flexibility, power consumption and suitability in the field of HPC and AI. FPGAs are claimed to have a highly flexible design which enables specific tuning of computationally intensive applications, making for high performance. On the other hand, GPUs have high parallel computing ability, and they are very apt in areas that involve a lot of parallelism, like deep learning. The framework used in this work includes the literature review, comparative analysis methodology, and the results section embellished with figures and tables. The results show that even though the GPUs are generally providing higher throughput necessary in many general-purpose AI applications, the FPGAs have certain advantages when it comes to power-constrained and application specific application requirements. This work ends with potential development and further possible areas of use of both technologies described above for the benefit of researchers and practitioners in the discipline.

References :

[1] Field-programmable gate array, Wikipedia. https://en.wikipedia.org/wiki/Field-programmable_gate_array

[2] Graphics processing unit, Wikipedia. https://en.wikipedia.org/wiki/Graphics_processing_unit#Sources

[3] FPGA, Microchipusa. https://www.microchipusa.com/manufacturer-articles/altera/alteras-max-10-fpga/

[4] Y. Wang, "Case Studies in HPC and AI: Comparing FPGA and GPU Performance," International Journal of High-Performance Computing Applications, vol. 34, no. 4, pp. 789-802, 2020.

[5] Tommason, GPU (Graphic Processing Unit), 2012. https://allyouneedtoknowict.wordpress.com/2012/10/22/gpu-graphic-processing-unit/

Keywords :

High-Performance Computing, Artificial Intelligence, FPGA, GPU, Programmability, Power Consumption, Parallel Processing, Energy Efficiency.