Sei sulla pagina 1di 9

International Journal of VLSI Design and Technology

Vol. 3: Issue 2
www.journalspub.com

Comparision of Adders for High Speed ALU


Rohit Kumar Bairwa*, Chanchal Soni, Roopendra Kumar, Deepak Kumar, Arun Dev Dhar
Dwivedi*
Department of Electrical and Electronic Engineering, Poornima University, Rajasthan, India

ABSTRACT
In this paper, logic optimized multiplexer based adders are incorporated in selected existing
adders like ripple carry adder, carry look-ahead adder, carry skip adder, carry select adder,
and carry save adder and its performance is analyzed in terms of area (slices used) and
maximum combinational path delay as a function of size. Here, different adder architectures
are simulated and analyzed based on power dissipation, area and speed by using Xilinx 14.4
ISE simulator and Verilog HDL. Hence, with the fast developments in the VLSI technology
and design, it is extremely important to include the interconnect sub circuitry during the
simulation process as efficiently as possible.

Keywords: carry look-ahead adder, carry skip adder, FPGA, full adder, half adder

*Corresponding Author
E-mail: adddwivedi@gmail.com, rohit.rohitb.bairwa11@gmail.com

INTRODUCTION Adder
A fast and accurate operation of digital In digital electronics adder is a digital
system is greatly influenced by the circuit that performs addition of numbers,
performance of resident adders. The most these can be classified into 1 bit adders
important for measuring the quality of and multi bit adders. Basic components
adder designs are computation time and
for an ALU are Adders in VlSI. There are
area. Adders are used not only in the
ALU(s), but also in other parts of the N number of adders each with their own
processor. advantages and disadvantages. When two
numbers are to be added and if each of
The basic principle in designing digital them is of N bits than we can add them in
adder circuit hovers around reducing the two different ways:
required hardware thus reducing the cost (a) Serial adders
too. To achieve this, logical optimization (b) Parallel adders
helps to obtaining minimum number of
literals to minimizing the transistor count Serial Adders
and the power consumption and increasing In serial addition the LSB's are added first
the speed of operation. It reduces circuit
than the carry created are propagated to
complexibility, so more flexible Circuit
constructed in processors. VLSI circuits the next higher bits. Serial adder is a
are available now in your computer, your digital device capable of adding two
car, your brand new state-of-the-art digital digital n-bit binary numbers, where n
camera, the cell-phones, and what have depends on the circuit implementation.
you. This technique helps in reducing the There are two types of adders which are
propagation delay and the power as follow:
consumption while maintaining low (a) Half adder
complexity of logic design. (b) Full adder

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 1


Adders for High Speed ALU Bairwa et al.

Half Adder and C IN inputs. We must also note that


A half adder is one of the most basic the C OUT will only be true if any of the
combinational logic circuits in computing. two inputs out of the three are HIGH.
The half adder produces two numbers as Boolean expression of sum and carry
output: the carry and the sum. The sum is output of full adder are,
the first digit of the answer, while the carry Sum = a XOR b
is any additional digit. It is called the carry Carry out = a AND b + b AND c + a AND c
because it carries over. Half adders only Equation for schematic full adder, are
work with single digit inputs. When O = ((I0 * I2) + (I1 * I2) + (I0 * I1));
numbers containing more than one digit
are to be combined together, a full adder is Parallel Adder
used. Parallel addition every it added in parallel
without waiting for carry and different
Boolean expression of sum and carry algorithms are used to compensate for the
output of half adder are: carry. There are different types of parallel
Sum = a XOR b adders and these are:
Carryout = a AND b (a) Ripple carry adder
(b) Carry look-ahead adder
Full Adder (c) Carry skip adder
The first two inputs are A and B and the (d) Carry select adder
third input is an input carry selected as C (e) Carry save adder
IN. By implementing the full-adder we
would be able to string up eight of them Ripple Carry Adder
composed to create a byte-wide adder and A ripple carry adder is a logic circuit in
cascade the carry bit from one adder to the which the carry-out of each full adder is
next. The output carry is designated as the carry in of the succeeding next most
Cout and the normal output is designated significant full adder. It is called a ripple
as S. carry adder because each carry bit gets
rippled into the next stage. In a ripple carry
In this truth table, we can see that the adder the sum and carry out bits of any of
output S is an EXOR between the input A the half adder stage is not valid until the
and the SUM of half-adder output with B carry occurs is of that stage.

Fig. 1. Half adder.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 2


International Journal of VLSI Design and Technology
Vol. 3: Issue 2
www.journalspub.com

Fig. 2. Full adder gate level block.

Fig. 3. RCA block.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 3


Adders for High Speed ALU Bairwa et al.

Fig. 4. RCA concept.

In simple words, the final result of the minimized when carry in to all stages are
ripple carry adder is valid only after the computed initially itself, as it will
joint propagation delays of all full adder minimize the wait for carry at every stage
circuits inside it. in a n bit adder, hence sum and carryout
expression can be re written as
Carry Look-Ahead Adder
As discussed in ripple carry adder, the Sum=(a)XOR(b)XOR(Cin)
significant delay produced due to ripple Carry out=(Gi)+(Pi).(Cin),
operation is a trade off. This can be Where, Pi=(a)XOR(b) and Gi=(a).(b).

Fig. 5. CLA.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 4


International Journal of VLSI Design and Technology
Vol. 3: Issue 2
www.journalspub.com

Carry Skip Adder For adding of two n-bit numbers with a


A carry-skip adder (also known as a carry- carry-select adder is completed with two
bypass adder) is an adder implementation adders (two ripple carry adders) in order to
that better on the delay of a ripple-carry perform the calculation twice, one time
adder with less effort compared to other with the assumption of the carry-in being
adders. The improvement in the worst-case zero and the other assuming it would be
delay is achieved by using different carry- one. After calculating both results, the
skip adders to form a block-carry-skip correct sum, and the correct carry-out, is
adder. In binary system, carry either can be then selected with the multiplexer once the
0 or 1. Only two possibility of value correct carry-in is known.
provides feasibility for choosing between
two. So, if there is a mechanism to select The general idea of the carry-select adder
carry, or better say, skipping carry through is to use blocks of two ripple-carry adders,
several stages, then delay minimization one of which is fed with a constant 0
can be better obtained. carry-in while the other is fed with a
constant 1 carry-in. Therefore, both blocks
Carry-skip-adders are chained (see block- can calculate in parallel.
carry-skip-adders) to reduce the overall
critical path. Carry Save Adder
A carry-save adder is a type of digital
Carry Select Adder adder, used in computer micro architecture
The carry-select adder basically consists of to compute the sum of three or more n-bit
two ripple carry adders and a multiplexer. numbers in binary.

Fig. 6. Carry skip logic block.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 5


Adders for High Speed ALU Bairwa et al.

Fig. 7. Carry select adder.

Fig. 8. Basic block carry save adder.

RESULT ANALYSIS verify and a synthesis report gives speed


The design of high speed ALU is and area of designed circuit. Hence, in this
necessary to increase the computation paper, the analysis of high speed adders
speed. The VLSI implementation of these physical and functional parameters like
high speed adders are done using RTL speed, delay, memory usage, number of
Codes(Verilog HDL) and simulation logic gates are measured. The results are
process done using Xilinx ISE simulator shown below.
14.4. The simulation process helps to

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 6


International Journal of VLSI Design and Technology
Vol. 3: Issue 2
www.journalspub.com

Comparison of Adders
4-bit adders
Table 1. Synthesis report of 4-bit adders.
S. no Parameters Ripple carry Carry look-ahead Carry skip Carry select
1 XOR 1 bit 4 8 8 8
2 NO. of Slices 4/960 4/960 8/4656 9/960
3 Levels of Logic 6 6 6 10
4 CPU Processing Time 17.33 sec 12.42 sec 15.87 sec 15.86 sec
5 Memory usage 250988 kb 251052 kb 250604 kb 251116 kb
6 Logic delay 6.723 ns 7.306 ns 7.306 ns 9.171 ns
7 Route delay 2.236 ns 2.576 ns 2.738 ns 4.032 ns
8 Total delay 8.959 ns 9.882 ns 10.044 ns 13.203 ns

8-bit adders
Table 2. Synthesis report of 8-bit adders.
S.no Parameters Ripple carry Carry look-ahead Carry skip Carry select
1 XOR 1 bit 8 16 16 24
2 NO. of slices 9/960 15/960 10/4656 18/960
3 Levels of logic 10 7 6 13
4 CPU Processing Time 17.85 sec 13.98 sec 9.09 sec 15.46 sec
5 Memory usage 251116 kb 251628 kb 251948 kb 252908 kb
6 Logic delay 9.171 ns 8.010 ns 9.923 ns 10.731 ns
7 Route delay 4.032 ns 3.305 ns 2.431 ns 5.163 ns
8 Total delay 13.203 ns 11.315 ns 9.356 ns 15.894 ns

16-bit adders
Table 3. Synthesis report of 16-bit adders.
S.no Parameters Ripple carry Carry look-ahead Carry skip Carry select
1 XOR 1 bit 16 32 32 24
2 NO. of Slices 18/960 32/960 16/4656 18/960
3 Levels of Logic 18 9 8 13
4 CPU Processing Time 16.20 sec 14.09 sec 9.46 sec 10.23 sec
5 Memory usage 252332 kb 253036 kb 252204 kb 252268 kb
6 Logic delay 14.067 ns 9.418 ns 8.331 ns 10.731 ns
7 Route delay 7.623 ns 4.781 ns 4.047 ns 5.163 ns
8 Total delay 21.690 ns 14.199 ns 12.378 ns 15.894 ns

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 7


Adders for High Speed ALU Bairwa et al.

CONCLUSION [5] S. Akashe, N.K. Tiwari, J. Shrivas,


In this paper, different type of adders like R. Sharma. A novel high speed &
ripple carry, carry select, carry skip, carry power efficient half adder design
look-ahead are simulated and synthesized using MTCMOS Technique in 45
on the platform Xilinx ISE simulator 14.4 nanometre regime, Advanced
and the parameters of these adders are Communication Control and
captured in synthesized report. Finally, the Computing Technologies
parameters like memory usage, levels of (ICACCCT), In: 2012 IEEE
logic, delay are compared for 4-bit, 8-bit International Conference.
and 16-bit adders. Ramanathapuram, 2012, 157–61p.
[6] V.G. Oklobdzija. High performance
From Table 3, this review paper concludes arithmetic units, In: High-
that the carry-skip adder is the efficient Performance System Design:Circuits
adder in comparision of these parameters, and Logic. Wiley-IEEE Press; 1999,
rather as we know that carry save adder is 405–28p.
most efficient than other adders but carry [7] S. Kim, M.C. Papaefthymiou. True
skip adder also have high efficient as single-phase energy-recovering logic
compared to ripple carry, carry select for low-power, high-speed VLSI,
adder, and carry look-ahead adder. Low Power Electronics and Design,
In: Proceedings 1998 International
REFERENCES Symposium. Monterey, CA, USA,
[1] P. Gurjar, R. Solanki, P. Kansliwal, 1998, 167–72p.
M. Vucha. VLSI implementation of [8] S. Kim, M.C. Papaefthymiou. True
adders for high speed ALU, In: single-phase energy-recovering logic
Annual IEEE India Conference. for low-power, high-speed VLSI,
Hyderabad, 2011, 1–6p. Low Power Electronics and Design,
[2] V. Kanimozhi, R. Gowri Shankar. In: Proceedings. 1998 International
Design and implementation of Symposium. Monterey, CA, USA,
Arithmetic Logic Unit using 1998, 167–72p.
modified novel bit adder in QCA, [9] W.F. Wallace, S.S. Dlay, O.R.
Innovations in Information, Hinton. Probabilistic carry state
Embedded and Communication estimate for improved asynchronous
Systems (ICIIECS), In: International adder performance, IEE Proc Comput
Conference. Coimbatore, 2015, 1–6p. Digital Tech. 2001; 148(6): 221–6p.
[3] R. Cherian, N. Thomas, Y. Shyju. [10] R.K. Kolagotla, H.R. Srinivas, G.F.
Implementation of binary to floating Burns. VLSI implementation of a
point converter using HDL, 200-MHz 16×16 left-to-right carry-
Automation, computing, free multiplier in 0.35 μm CMOS
communication, control and technology for next-generation DSPs,
compressed sensing (iMac4s), In: Custom Integrated Circuits
2013 International Multi-Conference. Conference, In: Proceedings of the
Kottayam, 2013, 461–4p. IEEE 1997. Santa Clara, CA, 1997,
[4] G. Jyothish Chandran, S.K. John. 469–72p.
Implementation of an area efficient [11] S. Khan, S. Kakde, Y. Suryawanshi.
data converter with increased VLSI implementation of reduced
effective number of bits, In: 2012 complexity wallace multiplier using
12th International Conference on energy efficient CMOS full adder,
Intelligent Systems Design and Computational Intelligence and
Applications (ISDA). Kochi, 2012, Computing Research (ICCIC), In:
662–7p.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 8


International Journal of VLSI Design and Technology
Vol. 3: Issue 2
www.journalspub.com

IEEE International Conference. Bipolar/BiCMOS Circuits and


Enathi, 2013, 1–4p. Technology Meeting, In:
[12] T. Harish Anand, D. Vaithiyanathan, Proceedings of the 1992.
R. Seshasayanan. Optimized Minneapolis, MN, 1992, 191–4p.
architecture for Floating Point [17] Y.-M. Huang, J.B. Kuo. A high-
computation Unit, Emerging Trends speed conditional carry select (CCS)
in VLSI, Embedded System, Nano adder circuit with a successively
Electronics and Telecommunication incremented carry number block
System (ICEVENT), In: (SICNB) structure for low-voltage
International Conference. VLSI implementation, In: IEEE
Tiruvannamalai, 2013, 1–5p. Transactions on Circuits and Systems
[13] B.K. Mohanty, P.K. Meher. VLSI II: Analog and Digital Signal
Architecture for high-speed/low- Processing. Vol. 47, no. 10, 2000,
power implementation of multilevel 1074–9p.
lifting DWT, APCCAS 2006 – 2006, [18] J.B. Kuo, H.J. Liao, H.P. Chen. A
In: IEEE Asia Pacific Conference on BiCMOS dynamic carry lookahead
Circuits and Systems. Singapore, adder circuit for VLSI
2006, 458–61p. implementation of high-speed
[14] B.K. Mohanty, P.K. Meher. Merged- arithmetic unit, IEEE J Solid-State
cascaded systolic array for VLSI Circ. 1993; 28(3): 375–8p.
implementation of discrete wavelet [19] C. Selvakumari, M. Jeyaprakash, A.
transform, APCCAS 2006 – 2006, In: Kavitha. Transistor level
IEEE Asia Pacific Conference on implementation of a 8 bit multiplier
Circuits and Systems. Singapore, using vedic mathematics in 180nm
2006, 462–5p. technology, In: 3rd International
[15] H. Park, et al. Design and Conference on Computing for
implementation and on-chip high- Sustainable Global Development
speed test of SFQ half-precision (INDIACom). New Delhi, India,
floating-point adders, IEEE Trans 2016, 1514–20p.
Appl Superconduct. 2009; 19(3): [20] M. Sangsefidi, M. Karimpour, M.
634–9p. Sarayloo. Efficient design of a
[16] J.B. Kuo, H.J. Liao, H.P. Chen. A coplanar adder/subtractor in
BiCMOS dynamic full adder circuit quantum-dot cellular automata, In:
for VLSI implementation of high- IEEE European Modelling
speed parallel multipliers using Symposium (EMS). Madrid, 2015,
Wallace tree reduction architecture, 456–61p.

IJVDT (2017) 1–9 © JournalsPub 2017. All Rights Reserved Page 9

Potrebbero piacerti anche