Doctor of Philosophy (PhD), Wright State University, 2022, Electrical Engineering
Binary multipliers are ubiquitous in digital hardware. Digital multipliers along with the adders play a major role in computing, communicating, and controlling devices. Multipliers are used majorly in the areas of digital signal and image processing, central processing unit (CPU) of the computers, high-performance and parallel scientific computing, machine learning, physical layer design of the communication equipment, etc. The predominant presence and increasing demand for low-power, low-cost, and high-performance digital hardware led to this work of developing optimized multiplier designs. Two optimized designs are proposed in this work. One is an optimized 8 x 8 Booth multiplier architecture which is implemented using 32nm CMOS technology. Synthesis (pre-layout) and post-layout results show that the delay is reduced by 24.7% and 25.6% respectively, the area is reduced by 5.5% and 15% respectively, the power consumption is reduced by 21.5% and 26.6% respectively, and the area-delay-product is reduced by 28.8% and 36.8% respectively when compared to the performance results obtained for the state-of-the-art 8 x 8 Booth multiplier designed using 32nm CMOS technology with 1.05 V supply voltage at 500 MHz input frequency. Another is a novel radix-8 structure with 3-bit grouping to reduce the number of partial products along with the effective
partial product reduction schemes for 8 x 8, 16 x 16, 32 x 32, and 64 x 64 signed multipliers. Comparing the performance results of the (synthesized, post-layout) designs of sizes 32 x 32, and 64 x 64 based on the simple novel radix-8 structure with the estimated performance measurements for the optimized Booth multiplier design presented in this work, reduction in delay by (2.64%, 0.47%) and (2.74%, 18.04%) respectively, and reduction in area-delay-product by (12.12%, -5.17%) and (17.82%, 12.91%) respectively can be observed. With the use of the higher radix structure, delay, area, and power consumption can be further reduced. (open full item for complete abstract)
Committee: Saiyu Ren Ph.D. (Advisor); Raymond E. Siferd Ph.D. (Committee Member); Henry Chen Ph.D. (Committee Member); Marian K. Kazimierczuk Ph.D. (Committee Member); Yan Zhuang Ph..D. (Committee Member); Michael Saville Ph.D., P.E. (Other); Barry Milligan Ph.D. (Other)
Subjects: Computer Engineering; Electrical Engineering