PERFORMANCE EVALUATION OF LOW POWER CARRY SAVE ADDER FOR VLSI APPLICATIONS

International Journal of VLSI design & Communication Systems (VLSICS) Vol.9, No.3, June 2018
DOI : 10.5121/vlsic.2018.9305 51
PERFORMANCE EVALUATION OF LOW POWER
CARRY SAVE ADDER FOR VLSI APPLICATIONS
Divya Tripathi and Subodh Wairya
Department of Electronics Engineering,
Institute of Engineering & Technology, Lucknow
ABSTRACT
This report examines the subject of sub threshold leakage on carry save adder. When the gate to source
voltage reduces to the threshold voltage at that place is yet some amount of current flow in the circuit and
that is undesired. As the process technology advancing much rapidly the threshold voltage of MOS devices
reduces very drastically, and it must be applied in lower power devices since it contributes to low amount
of leakage current which confine increases the power consumption of the devices. Adders are the basic
building blocks for any digital circuit design and used in almost all arithmetic’s. The CSA proves efficient
adders due to its quick and precise computations. Hence this paper performs sub threshold analysis on
CSA and the scrutinize results that the total average power is around 4.93µW, the propagation delay for
complete operation is 16.3ns and since this design uses GDI cell so there is a reduction in area with 37%.
.
KEYWORDS
Sub threshold Leakage, Gate Diffusion Input (GDI), Carry Save Adder (CSA), Leakage current, Transistor
Modeling.
1. INTRODUCTION
Arithmetic circuits are basic building blocks of any digital IC design. The major offices such as
summation, subtraction, multiplication, and variance are used oftentimes in an arithmetic and
logic unit of any digital IC. Since the addition and multiplication are nearly shared similar
functionality so it is possible to design adders which perform both the functions very efficiently
[1]. A carry save adder (CSA) is distinguished from other types of adders by the fact that the
carry bits and half-sum bits which result from each addition are not immediately combined or
consolidated but instead are saved separately from each other for subsequent use in the next
addition performed by the CSA, the inputs to which will comprise the saved carry and half-sum
bits (the latter hereinafter referred to simply as "sum bits" for convenience) and the bits of an
operand which is being added to, or in some instances being effectively subtracted from, the value
jointly represented by these saved carry and sum bits.
Carry save adders commonly are employed in high-speed multipliers, where they generally are
able to function more rapidly than "carry propagate" or "ripple carry" adders because a carry save
adder does not completely perform the relatively time-consuming process of combining carries
with sum bits between successive additions in the multiplication process but instead defers this
task until the final cycle of the multiplying operation. This work design the CSA in such a manner
that it should consume less silicon area and perform efficient operations; to achieve this the GDI
methodology is used to design the 4-bit CSA. The subthreshold leakage of CSA is measured and
optimized by adjusting the MOS process parameters such as oxide thickness, channel
length/width, junction depth and so forth reduce the leakage current of the device [4]. The basic

52
logical operation of 4-bit CSA can be realized by the following fig 1. By using CSA not only the
focal ratio of the computation increases, but as well it sustains the power since the carry
propagation is saved between previous and next state if generated any In this paper a 4-bit CSA is
designed using GDI methodology and its performance is examined with respect to subthreshold
leakage.
This work is organized in the following sections: Section 2 provides a necessary existing work
related to the CSA circuit. This section also deals an existing work with its pros and cons. Section
3 details the architecture, including the CSA design and explication of results. Section 4 provides
the comparative results based on prior work and section 5 their conclusion
Fig. 1 Logical Operation of 4-Bit CSA
2. LITERATURE SURVEY
Tooraj Nikoubin [3] proposed three inputs xor/xnor circuit using a systematic cell design
methodology (SCDM) in hybrid CMOS logic style. Their design shows, full swing and balanced
outputs; even with supply voltage scaling they perform well and in the critical path they contain
two transistors. It was shown that the proposed circuits exhibit 27% - 77% decrease in average
energy delay product and also 26% - 32% improvement in the field.
Ramyanshu Datta [6] present a 4-to-2 CSA using dynamic logic and limited switch dynamic
logic (LSDL) circuit family. The operation of the adder is improved as per the results. Further,
they used a latching element which contains the ability of dynamic circuit. The LSDL
methodology shows significant improvement with regard to power dissipation, leakage and area.
Adarsh Kumar Agarwal, S.Wairya [7] employed high speed and low power performance of single
GDI adder, its speed performance is degraded when cascaded within a long chain because it is
equivalent to a RC ladder network. As it is well known that the delay of single topology chain is a
quadratic function of the number of Full Adders so for the long single topology chain the delay
becomes unacceptably high. To limit this problem mixed topology strategy is adopted. This
strategy includes the interruption of inverters (static gate) within the chain. proposed GDI Full
Adder based on inverter is a better one as compared to existing topologies as far as delay, power
and PDP is concerned. On the basis of simulation results obtained it is culminated that the
proposed topology is suitable for high speed low power arithmetic circuits.
R. Mahalakshmi and T. Sasilatha [9] made some major advances in the design of CSA using
CMOS technology. This works mainly focus on power using up and overall chip area since for an
older structure these are the two primary parameters to be worried about. Their design uses
250nm and 65nm process technology, respectively, and their report shows a good quantity of
reduction in power and expanse.

53
Amuthavalli. G and Gunasundari. R [12] proposed a power aware design for ripple carry adder
(RCA). The proposed circuit also helps in cutting down the leakage current of the adder. This paper
also analyzes the subthreshold leakage and it was proven that power due to subthreshold leakage
reduces significantly as compared to conventional RCA design. They also employ the multi-
Threshold CMOS (MT-CMOS) in the design for maintaining low power performance.
Ravikumar A. Javeli [15] design, CSA by using carry look ahead adder (CLA) instead of RCA,
and the design approach represents the improvement in speed of the increase. The proposed
design compared with regard to area, power and timings and it was established that the CSA using
CLA improves the speed of calculation with minimum effect on area and the power of the pattern
cell.
3. PROPOSED ARCHITECTURE
As MOS integrated circuit technology has evolved to exploit smaller and smaller device
structures, it has become increasingly important in recent years to look more closely at the
minority carriers present under the gate when the gate voltage is less than threshold, i.e. in what is
called the “sub-threshold” region. These carriers cannot be totally neglected, and play an
important role in device and circuit performance. At first they were viewed primarily as a
problem, causing undesirable “leakage” currents and limiting circuit performance. Now it is
recognized that they also enable a very useful mode of MOSFET operation, and that the sub
threshold region of operation is as important as the traditional cut-off, linear, and saturations
regions of operation. [3]-[5].
It consists of seven full adders and single half adder. The various ties and signal flow are shown
in the design. The transistors cells are utilized in the design of CSA are GDI cells. GDI method is
a low power design technique, by using this technique the various parameters such as power
waste, delay, and sweep of digital IC is reduced much extent with very less complexity in the
logic pattern. 4-bit CSA block diagram is shown in fig. 2. Circuit diagram of 4-bit CSA with
equivalent GDI cells is shown in fig. 3. A proposed 4-bit full adder using GDI cell is shown in
fig.4. This carry save adder uses only 10 transistors to perform both sum and carry operations.
Fig.2. 4-bit CSA Block Diagram

54
Fig.3. Circuit diagram of 4-bit CSA with equivalent GDI cells
Fig.4. 4-bit CSA Block Diagram
4. SIMULATIONS AND RESULTS
The proposed carry save adder design is simulated using HSPICE and by taking 32nm CMOS
nano technology. The functionality of the CSA is verified by writing HDL code. The mentor
graphics modelsim used for the HDL modeling of the proposed CSA. The design of CSA is done
using three level hierarchy, i.e. first the basic cell is designed and then using that cells the various
subsystems are designed, and finally using these subsystems 4- bit CSA is designed. To design
subsystems the GDI cells are applied to cut the power and area of the device. The pulse signal is
chosen for input to the CSA. Fig.5 shows the input signals to the CSA.

55
Fig.5. Input waveform of CSA
The output simulation results demonstrate the proper operation of the 4-bit CSA, and the behavior
of various signals involved in the design is expressed beneath. The fig. 6 shows the output
response of the CSA. HDL simulation can be done only to accomplish a high degree of trust in its
correctness before commencing design, and may simulate one to one orders of magnitude quicker
than a gate level description. An HDL specification for a part can perform the basis for a
simulation model to verify the operation of the part in the wider system context; this depends on
how accurately the specification handles aspects such as timing and initialization. The fig.7 shows
the HDL model of CSA.
Fig.6. Output waveform from CSA

56
Fig.7. HDL Model of CSA
The fig. 8 shows the DC characteristics of the CSA. Fig. 9 and 10 shows the relationship between
Ids vs Vds and Ids vs Vgs respectively. Granting to the DC characteristic it can be reasoned that the
MOS is not biased from its normal operation and the subthreshold effect is also negligible so the
low power operation is required.
Fig.8. DC Characterstics of CSA Fig.9. Ids vs. Vds
Fig.10. Ids Vs Vgs Fig.11. VT wrt Channel Length
Fig.11 shows the relationship between channel length and the threshold voltage, after a certain
limit is turned over for short channel the threshold roll off is well determined from the graph
which will serve in the deciding the proper channel length of the device and it improves the
device operation. The various optimized parameters related to the simulation are listed in Table –
1.Fig.12 shows the two major parasitic capacitances of the one bit full adder. However, these
capacitances are optimized in such a fashion that it will not affect the circuit performance. The
final optimized simulation layout is shown in figure – 13.

57
Table.1.Result of proposed design
Fig.12. Parasitic Capacitances Fig.13. Optimized Simulation Layout
5 CONCLUSION
The 4-bit CSA is designed and related waveforms are plotted. By optimizing the operation
parameters, it is noted that the CSA performs as it should and in subthreshold it consumes very
less power since its static leakage current is cut. The switching also reduced by reducing carries
propagation to the subsequent stages which ultimately result in a decrease of power dissipation of
the gimmick. The results also bring out that the GDI cell requires less area as compared to
conventional CMOS and it also improves the speed of operation of the device. Hence this paper
performs sub threshold analysis on CSA and the scrutinize results that the total average power is
around 4.93µW, the propagation delay for complete operation is 16.3ns and since this design uses
GDI cell so there is a reduction in area with 37%.
REFERENCES
[1] K. Granhaug and S. Aunet, (2006), “ Six sub-threshold full adders cells characterized in 90 nm
CMOS technology, “ IEEE Workshop On Design And Diagnostics Of Electronic Circuits And
System, pp.25-30.
[2] H.Thap, liyal, N.Gopi, K.K.P. Kumar, (2006),“Low Power Hierarchical Multiplier And Carry look-
ahead Architecture”,IEEE International Conference On Computer Systems And Applications, pp.88-
92.
[3] Tooraj Nikoubin, Mahdieh Grailoo, and Changzhi Li, (2016) “Energy and Area Efficient Three-Input
XOR/XNORs with Systematic Cell Design Methodology”, IEEE Transactions on Very Large Scale
Integration (VLSI) Systems, Volume: 24, Issue: 1, pp. 398 - 402.

58
[4] Manash Chanda, Sankalp Jain, Swapnadip De, Chandan Kumar Sarkar, ( 2015) “Implementation of
Subthreshold Adiabatic Logic for Ultralow-Power Application”, IEEE Transactions on Very Large
Scale Integration (VLSI) Systems, Volume: 23, Issue: 12, pp. 2782 – 2790.
[5] H.Mahmoodi-Meimand, A.Afzali-Kusha and M.Nourani, ( 2001) “Adiabatic carry look-ahead adder
with efficient power clock generator”, IEEE Proceedings - Circuits, Devices and Systems, Volume:
148, Issue: 5, pp. 229 – 234.
[6] Ramyanshu Datta, J. A. Abraham, Robert Montoye, Wendy Belluomini, Hung Ngo, Chandler
McDowell, JB Kuang, Kevin Nowk (2004) “A Low Latency and Low Power Dynamic Carry Save
Adder”, Proceedings of the International Symposium on Circuits and Systems, vol.2, pp.477-480
[7] Adarsh Kumar Agarwal, S.Wairya, R.K. Nagaria and S. Tiwari (2009) “A New Mixed Gate Diffusion
input Full Adder Topology for High Speed Low Power Digital Circuits,” World Applied Sciences
Journal, IDOSi Publication, vol.7, pp.138-144
[8] Yanxiang Huang, Ajay Kapoor , Robert Rutten and Jose Pineda de Gyvez ( 2013) “A 13 bits 4.096
GHz 45 nm CMOS Digital Decimation Filter Chain Using Carry-Save Format Numbers”,
NORCHIP-13, 978-1-4799-1674-4/13.
[9] R.Mahalakshmi and T.Sasilatha,(2013) “A power efficient carry save adder and modified carry save
adder using CMOS technology”, IEEE International Conference on Computational Intelligence and
Computing Research (ICCIC), 978-1-4799-1597-2
[10] S. Sakthikumaran, S. Shalivahnan, V. S. Kanchana Bhaskara, V. Kavinilavu, B. Vrintha, C. Vinod,
(2011) “A Very Fast and Low Power Carry Select Adder Circuit”, 3rd International Conference on
Electronics Computer Technology (ICECT), vol.3, pp.147-151.
[11] Dilip Kumar, Gayen, Jitendra Nath Roy, and Rajat Kumar Pal,(2012) “All-Optical Carry Look Ahead
Adder with the Help of Terahertz Optical Asymmetric De multiplexer”, Volume 123, Issue 1, Pages
40–45.
[12] Amuthavalli. G and Gunasundari R,(2015) “Analysis and Design of Subthreshold Leakage Power-
Aware Ripple Carry Adder at Circuit-Level using 90nm Technology”, Volume 48, Pages 660-665.
[13] S.Wairya, R.K. Nagaria, S. Tiwari, (2012) “Performance Analysis of High Speed Hybrid CMOS Full
Adder Circuits for Low Voltage VLSI Design”, Hindawi Publishing Corporation VLSI Design,
Volume 2012, Article ID 173079.
[14] S. Wairya, H. Pandey, R.K. Nagaria, S. Tiwari, (2011) “Ultra Low Voltage High Speed 1-Bit
CMOSAdder”, 10.1109/ ICPCES.2010.5700479.
[15] Ravikumar A Javali, Ramanath J Nayak, Ashish M Mhetar, Manjunath C Lakkannavar,(2014)
“Design of High Speed Carry Save Adder using Carry Look ahead Adder”, International Conference
on Circuits, Communication, Control and Computing (I4C) vol.2, pp163-172.
AUTHOR’S PROFILE
Divya Tripathi she did B. Tech. in Electronics & Communication from College of
Engineering & Technology, Lucknow, India in 2009 and M. Tech in VLSI Design in
from Jayoti Vidyapeeth Women’s University, Jaipur, India, in 2012. Currently
pursuing her Ph.D. Her current research interests include low power, high speed VLSI
circuits.

PERFORMANCE EVALUATION OF LOW POWER CARRY SAVE ADDER FOR VLSI APPLICATIONS

Recommended

More Related Content

What's hot (18)

Similar to PERFORMANCE EVALUATION OF LOW POWER CARRY SAVE ADDER FOR VLSI APPLICATIONS (20)

Recently uploaded (20)

PERFORMANCE EVALUATION OF LOW POWER CARRY SAVE ADDER FOR VLSI APPLICATIONS