The Journal of the Korea institute of electronic communication sciences (한국전자통신학회논문지)

Korea Institute of Electronic Communication Science (한국전자통신학회)
권호 표지
DOI QR코드

DOI QR Code

Design of 32-bit Floating Point Multiplier for FPGA

FPGA를 위한 32비트 부동소수점 곱셈기 설계

  • ;
  • 김대익 (전남대학교 전자통신공학과)
  • Received : 2024.01.16
  • Accepted : 2024.04.12
  • Published : 2024.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

With the expansion of floating-point operation requirements for fast high-speed data signal processing and logic operations, the speed of the floating-point operation unit is the key to affect system operation. This paper studies the performance characteristics of different floating-point multiplier schemes, completes partial product compression in the form of carry and sum, and then uses a carry look-ahead adder to obtain the result. Intel Quartus II CAD tool is used for describing Verilog HDL and evaluating performance results of the floating point multipliers. Floating point multipliers are analyzed and compared based on area, speed, and power consumption. The FMAX of modified Booth encoding with Wallace tree is 33.96 Mhz, which is 2.04 times faster than the booth encoding, 1.62 times faster than the modified booth encoding, 1.04 times faster than the booth encoding with wallace tree. Furthermore, compared to modified booth encoding, the area of modified booth encoding with wallace tree is reduced by 24.88%, and power consumption of that is reduced by 2.5%.

빠른 고속 데이터 신호 처리 및 논리 연산을 위한 부동 소수점 연산 요구 사항이 확대됨에 따라 부동 소수점 연산 장치의 속도는 시스템 작동에 영향을 미치는 핵심 요소이다. 본 논문에서는 다양한 부동소수점 곱셈기 방식의 성능 특성을 연구하고, 캐리와 합의 형태로 부분 곱을 압축한 다음, 최종 결과를 얻기 위해 캐리 미리 보기 가산기를 사용한다. Intel Quartus II CAD 툴을 이용하여 Verilog HDL로 부동소수점 곱셈기를 기술하고 성능 평가를 하였다. 설계된 부동소수점 곱셈기는 면적, 속도 및 전력 소비에 대해 분석 및 비교하였다. 월러스 트리를 사용한 수정 부스 인코딩 방식의 FMAX는 33.96Mhz로 부스 인코딩보다 2.04배, 수정 부스 인코딩보다 1.62배, 월러스 트리를 사용한 부스 인코딩보다 1.04배 빠르다. 또한, 수정 부스 인코딩에 비해 월러스 트리를 이용한 수정 부스 인코딩 방식의 면적은 24.88% 감소하고, 전력소모도 2.5% 감소하였다.

Keywords

References

  1. S. Kim, H. Seo, S. Kim, and D. Kim, "Approximate Multiplier With Efficient 4-2 Compressor and Compensation Characteristic," J. of the Korea Institute of Electronic Communication Sciences, vol. 17, no. 1, 2022, pp. 173-180.
  2. H. Seo and D. Kim, "Approximate multiplier with high density, low power and high speed using efficient partial product reduction," J. of the Korea Institute of Electronic Communication Sciences, vol. 17, no. 4, 2022, pp. 671-678.
  3. J. Kim and S. Lee, "High-Performance Multiplier Using Modified Gate-Diffusion Input (m-GDI) Compresso," J. of the Korea Institute of Electronic Communication Sciences, vol. 18, no. 2, 2023, pp. 285-290.
  4. A. Akkas and M. J. Schulte, "A quadruple precision and dual double precision floating-point multiplier," In Proc. Euromicro. Symp. on Digital System Design, Belek-Antalya, Turkey, Sept. 2003, pp. 76-81.
  5. S. V. Siddamal, R. M. Banakar, and B. C. Jinaga, "Design of High-Speed Floating Point Multiplier," In Proc. 4th IEEE Int. Symp. on Electronic Design, Test and Applications (delta 2008), Hong Kong, China, Jan. 2008, pp. 285-289.
  6. IEEE Std. 754-1985, IEEE Standard for Binary Floating-point Arithmetic. IEEE, New York, NY, 1985.
  7. T. Krishnan and S. Saravanan, "Design of Low-Area and High Speed Pipelined Single Precision Floating Point Multiplier," In Proc. 2020 6th Int. Conf. (ICACCS). Communications, Coimbatore, India, Mar. 2020, pp. 1259-1264.
  8. Y. Guo, H. Sun, and S. Kimura, "Small-Area and Low-Power FPGA-Based Multipliers using Approximate Elementary Modules," In Proc. 2020 25th Asia and South Pacific Design Automation Conf. (ASP-DAC). Communications, Beijing, China, 2020, pp. 599-604.
  9. X. Jiang, P. Xiao, M. Qiu, and G. Wang, "Performance effects of pipeline architecture on an FPGA-based binary 32 floating point multiplier," J. of Microprocessors and Microsystems, vol. 37, no. 8, 2013, pp. 1183-1191. https://doi.org/10.1016/j.micpro.2013.08.007
  10. M. J. Rao and S. Dubey, "A high speed and area efficient Booth recoded Wallace tree multiplier for fast arithmetic circuits," In Proc. 2012 Asia Pacific Conf. on Postgraduate Research in Microelectronics and Electronics, Communications, Hyderabad, India, Dec. 2012, pp. 220-223.
  11. Z. Zhao and Z. Line, "Design of an improved Wallace tree multiplier," J. of Electronic Design Applications, no. 8, 2006, pp. 113-116.
  12. S. Yuan and C. Zhang, "A Design of High-Speed 4-2 Compressor for Fast Multiplier," J. of Microelectronics & Computer, vol. 19, no. 4, 2002, pp. 53-56.
  13. K. Thiruvenkadam, J. Ramesh, and A. S. Pillai, "Area-efficient dual-mode fused floating-point three-term adder," J. of Circuits, Systems, and Signal Processing, vol. 38, no. 1, 2019, pp.173-190.  https://doi.org/10.1007/s00034-018-0848-y