# Datapath Blocks

- 1.4k Downloads

## Abstract

This chapter presents the design of the first three ultra-low-voltage prototypes which have been implemented in this book. These prototypes all consist of datapath blocks. Their target was to be able to operate at ultra-low supply voltages, while achieving high energy-efficiency, a speed of *n* × 10MHz and a high yield through variation-resilience. This chapter builds further upon the conclusions of the analyses of different gate-level building blocks and of architectural design choices. The chapter starts with a discussion on the design of the first prototype, which is a 32-bit logarithmic adder fabricated in a 90nm CMOS technology. This prototype is employed as a proof of concept to confirm the robust operation of Transmission Gate logic and latch-based deep pipelining in the ultra-low-voltage region. Extensive measurement results evaluate their successful functionality and are compared to the state-of-the-art. Furthermore, this chapter examines the ultra-low-voltage design of a 16-bit Multiply-Accumulate unit. This MAC has been fabricated in both the 90nm and the 40nm CMOS technologies at hand, resulting in the second and the third prototype. Gate-level and architectural improvements with respect to the design of the adder are implemented and tested. An extensive comparison between the measurement results of the MAC in both CMOS technologies allows studying the impact of scaling on ultra-low-voltage designs.

## Keywords

Datapath Block Multiply-accumulate Unit 90-nm CMOS Technology Resilience Variables Architectural Design Choices## References

- 1.Bol D, Ambroise R, Flandre D, Legat JD (2009) Interests and limitations of technology scaling for subthreshold logic. IEEE Trans Very Large Scale Integ (VLSI) Syst 17(10):1508–1519. DOI:10.1109/TVLSI.2008.2005413CrossRefGoogle Scholar
- 2.Bol D, Kamel D, Flandre D, Legat JD (2009) Nanometer MOSFET effects on the minimum-energy point of 45nm subthreshold logic. In: Proceedings of the ACM/IEEE international symposium on low power electronics and design (ISLPED), pp 3–8. DOI:10.1145/1594233.1594237Google Scholar
- 3.Brent RP, Kung HT (1982) A regular layout for parallel adders. IEEE Trans Comput C-31(3):260–264. DOI:10.1109/TC.1982.1675982CrossRefMathSciNetGoogle Scholar
- 4.Calhoun B, Chandrakasan A (2006) Ultra-dynamic voltage scaling (UDVS) using sub-threshold operation and local voltage dithering. IEEE J Solid State Circuits 41(1):238–245. DOI:10.1109/JSSC.2005.859886CrossRefGoogle Scholar
- 5.Hatamian M, Cash G (1986) A 70-MHz 8-bitx8-bit parallel pipelined multiplier in 2.5-
*μ*m CMOS. IEEE J Solid State Circuits 21(4):505–513. DOI:10.1109/JSSC. 1986.1052564CrossRefGoogle Scholar - 6.Kao J, Miyazaki M, Chandrakasan A (2002) A 175-mV multiply-accumulate unit using an adaptive supply voltage and body bias architecture. IEEE J Solid State Circuits 37(11):1545–1554. DOI:10.1109/JSSC.2002.803957CrossRefGoogle Scholar
- 7.Kogge PM, Stone HS (1973) A parallel algorithm for the efficient solution of a general class of recurrence equations. IEEE Trans Comput C-22(8):786–793, DOI:10.1109/TC.1973.5009159CrossRefMathSciNetGoogle Scholar
- 8.Reynders N, Dehaene W (2011) A 190mV supply, 10MHz, 90nm CMOS, pipelined sub-threshold adder using variation-resilient circuit techniques. In: Proceedings of the IEEE Asian solid-state circuits conference (A-SSCC), pp 113–116. DOI:10.1109/ASSCC.2011.6123617Google Scholar
- 9.Reynders N, Dehaene W (2012) Variation-resilient sub-threshold circuit solutions for ultra-low-power digital signal processors with 10MHz clock frequency. In: Proceedings of the IEEE European solid-state circuits conference (ESSCIRC), pp 474–477. DOI:10.1109/ESSCIRC. 2012.6341358Google Scholar
- 10.Reynders N, Dehaene W (2015) On the effect of technology scaling on variation-resilient sub-threshold circuits. Elsevier Solid State Electron 103:19–29CrossRefGoogle Scholar
- 11.Tajalli A, Leblebici Y (2011) Design trade-offs in ultra-low-power digital nanoscale CMOS. IEEE Trans Circuits Syst Regul Pap 58(9):2189–2200. DOI:10.1109/TCSI.2011.2112595CrossRefMathSciNetGoogle Scholar