Method of creation of FPGA based implementation of artificial intelligence as a service

Artem Perepelitsyn

Abstract


The subject of study in this article is the technologies of Field Programmable Gate Array (FPGA), methods, and tools for prototyping of hardware accelerators of Artificial Intelligence (AI) and providing it as a service. The goal is to reduce the efforts of creation and modification of FPGA implementation of Artificial Intelligent projects and provide such solutions as a service. Task: to analyze the possibilities of heterogeneous computing for the implementation of AI projects; analyze advanced FPGA technologies and accelerator cards that allow the organization of a service; analyze the languages, frameworks, and integrated environments for the creation of Artificial Intelligence projects for FPGA implementation; propose a technique for modifiable FPGA project prototyping to ensure a long period of compatibility with integrated environments and target devices; propose a technique for the prototyping of FPGA services with high performance to improve the efficiency of FPGA based AI projects; propose a sequence of optimization of neural networks for FPGA implementation; and provide an example of the practical implementation of the research results. According to the tasks, the following results were obtained. Analysis of the biggest companies and vendors of FPGA technology is performed. Existing heterogeneous technologies and potential non-electronic mediums for AI computations are discussed. FPGA accelerator cards with a large amount of High Bandwidth Memory (HBM) on the same chip package for implementation of AI projects are analyzed and compared. Languages, frameworks, and technologies as well as the capabilities of libraries and integrated environments for prototyping of FPGA projects for the AI applications are analyzed in detail. The sequence of prototyping of FPGA projects that are stable to changes in the environment is proposed. The sequence of prototyping of highly efficient pipelined projects for data processing is proposed. The steps of optimization of neural networks for FPGA implementation of AI applications are provided. An example of practical use of the results of research, including the use of sequences is provided. Conclusions. One of the main contributions of this research is the proposed method of creation of FPGA based implementation of AI projects in the form of services. Proposed sequence of neural network optimization for FPGA allows the reduction of the complexity of the initial program model by more than five times for hardware implementation depending on the required accuracy. The described solutions allow the construction of completely scalable and modifiable FPGA implementations of AI projects to provide it as a service.

Keywords


FPGA; FPGA-as-a-Service; FaaS; Artificial Intelligence; Artificial Intelligence-as-a-Service; AIaaS; hardware implementation of AI; heterogeneous computing for AI; neural network optimization; DPU

Full Text:

PDF

References


Perepelitsyn, A., Kulanov, V., & Zarizenko, I. Method of QoS evaluation of FPGA as a service. Radioelectronic and Computer Systems, 2022, no. 4, pp. 153–160. DOI: 10.32620/reks.2022.4.12.

Perepelitsyn, A., & Kulanov, V. Technologies of FPGA-based projects Development Under Ever-changing Conditions, Platform Constraints, and Time-to-Market Pressure. Proceedings 2022 IEEE 12th International Conference on Dependable Systems, Services and Technologies, DESSERT 2022, 2022, pp. 1-5, DOI: 10.1109/DESSERT58054.2022.10018828.

Hu, N., Wang, C., & Zhou, X. FLIA: Architecture of Collaborated Mobile GPU and FPGA Heterogeneous Computing. Electronics, 2022, vol. 11, iss. 22, no. 3756. pp. 1-14. DOI: 10.3390/electronics11223756.

Liang, B., Wang, S., Huang, Y., Liu, Y., & Ma, L. F-LSTM: FPGA-Based Heterogeneous Computing Framework for Deploying LSTM-Based Algorithms. Electronics, 2023, vol. 12, iss. 5, no. 1139. pp. 1-15. DOI: 10.3390/electronics12051139.

How does Let’s Enhance increases image resolution? Available at: https://help.letsenhance.io/en/ article/how-does-lets-enhance-increases-image-resolution-3p5z0p/ (accessed February 28, 2023).

Maksymov, I. Analogue and Physical Reservoir Computing Using Water Waves: Applications in Power Engineering and Beyond. Energies, 2023, vol. 16, iss. 14, no. 5366. pp. 1-26. DOI: 10.3390/en16145366

Khoram, E., Chen, A., Liu, D., Ying, L., Wang, Q., Yuan, M., & Yu, Z. Nanophotonic media for artificial neural inference. Photon. Res. 2019, vol. 7, iss. 8, pp. 823-827. DOI: 10.1364/PRJ.7.000823.

Alveo Product Selection Guide, Data Center Accelerator Cards, Xilinx. Available at: https://www.xilinx.com/content/dam/xilinx/support/documents/selection-guides/alveo-product-selection-guide.pdf. (accessed February 28, 2023).

Alveo U55C Data Center Accelerator Cards Data Sheet, AMD, DS978 (v1.3). Available at: https://docs.xilinx.com/r/en-US/ds978-u55c. (accessed June 23, 2023).

VHK158 Evaluation Board User Guide, AMD, UG1611 (v1.0). Available at: https://docs.xilinx.com/r/en-US/ug1611-vhk158-eval-bd. (accessed July 18, 2023).

Alveo UL3524 Ultra Low Latency Trading Data Sheet, AMD, DS1009 (v1.1). Available at: https://docs.xilinx.com/r/en-US/ds1009-ul3524. (accessed September 27, 2023).

Barkovska, O., Filippenko, I., Semenenko, I., Korniienko, V., & Sedlaček, P. Adaptation of FPGA architecture for accelerated image preprocessing. Radioelectronic and Computer Systems, 2023, no. 2, pp. 94-106. DOI: 10.32620/reks.2023.2.08

Zynq DPU Product Guide, Xilinx, PG338 (v3.3). Available at: https://docs.xilinx.com/r/3.3-English/pg338-dpu. (accessed February 28, 2023).

Perepelitsyn, A., Fesenko, H., Kasapien, Y., & Kharchenko, V. Technological Stack for Implementation of AI as a Service based on Hardware Accelerators. Proceedings 2022 IEEE 12th International Conference on Dependable Systems, Services and Technologies, DESSERT 2022, 2022, pp. 1-5. DOI: 10.1109/DESSERT58054.2022.10018615.

UltraFast Design Methodology Guide for Xilinx FPGAs and SoCs, Xilinx, UG949 (v2021.2). Available at: https://docs.xilinx.com/r/2021.2-English/ug949-vivado-design-methodology/SLR-Utilization-Considerations. (accessed February 28, 2023).

Vivado Design Suite Properties Reference Guide, Xilinx, UG912 (v2022.1). Available at: https://docs.xilinx.com/r/2022.1-English/ug912-vivado-properties/USER_CROSSING_SLR. (accessed February 28, 2023).

Perepelitsyn, A., Kasapien, Y., Fesenko, H., & Kharchenko, V. Technologies for Implementing of Artificial Intelligence as a Service based on Hardware Accelerators. Aerospace Technic and Technology, 2022, no. 6, pp. 57-65. DOI: 10.32620/aktt.2022.6.07.




DOI: https://doi.org/10.32620/reks.2023.3.03

Refbacks

  • There are currently no refbacks.