DESIGN AND IMPLEMENTATION OF LOW POWER ALU USING CLOCK GATING AND CARRY SELECT ADDER

http://www.iaeme.com/IJECET/index.asp 116 editor@iaeme.com
International Journal of Electronics and Communication Engineering and Technology
(IJECET)
Volume 7, Issue 3, May–June 2016, pp. 116–129, Article ID: IJECET_07_03_014
Available online at
http://www.iaeme.com/IJECET/issues.asp?JType=IJECET&VType=7&IType=3
Journal Impact Factor (2016): 8.2691 (Calculated by GISI) www.jifactor.com
ISSN Print: 0976-6464 and ISSN Online: 0976-6472
© IAEME Publication
DESIGN AND IMPLEMENTATION OF LOW
POWER ALU USING CLOCK GATING AND
CARRY SELECT ADDER
J. Sneha
Department of ECE, AITAM, Tekkali, India
Dr. A. S. Srinivasa Rao
A. Rajesh
ABSTRACT
CPUs in general-purpose personal computers, such as desktops and
laptops, dissipate significantly more power in the order of few watts because
of their higher complexity and speed. ALU is a fundamental building block of
CPU. It does all process related to arithmetic and logic operations. As the
operations become more complex, the ALU become more complex, more
expensive, takes up more space in the CPU and contributes more power
dissipation within the CPU. Hence power consumption of ALU is a major
issue in the designing of CPU.
In this paper, it is proposed to design an ALU with latch free clock gating
technique to reduce clock power and dynamic power by turning off unused
areas while working on one area of design to achieve low power consumption
and the results are compared with ALU without clock gating. And also further
improvement in power is expected by using a carry select adder (CSLA) as the
primary computational element of the arithmetic unit for fast arithmetic
operations. From the structure of the CSLA, it is clear that there is scope for
reducing the area and power consumption in the CSLA. This work uses a
simple and efficient gate-level modification to significantly reduce the area
and power of the CSLA. Based on this modification 16-bit square root CSLA
(SQRT CSLA) architecture have been developed and compared with the
regular SQRT CSLA architecture. The proposed design has reduced area and
power as compared with the regular SQRT CSLA with only a slight increase
in the delay. In this paper a low power 16 bit ALU is designed with HDL
programming and then implemented on Xilinx Spartan 3E FPGA.

Design and Implementation of Low Power Alu Using Clock Gating and Carry Select Adder
Key words: Low Power, ALU, Dynamic Power, Clock Power, Clock Gating,
Carry Select Adder (CSLA), FPGA, HDL Programming.
Cite this Article: J. Sneha, Dr.A.S.Srinivasa Rao and A.Rajesh, Design and
Implementation of Low Power Alu Using Clock Gating and Carry Select
Adder, International Journal of Electronics and Communication Engineering
& Technology, 7(3), 2016, pp. 116–129.
http://www.iaeme.com/IJECET/issues.asp?JType=IJECET&VType=7&IType=3
1. INTRODUCTION
Day by day the advancement in technology leads to increase in the number of
transistor count on any digital logic design .This increase in transistor count have an
sever impact on power consumption because adding more and more transistor will
give rise in the heat dissipated in the device [1]. Now a days the era is revolving
around portability, is offered by low power consuming system ,Since most of the
portable devices are battery driven the power consumption of these devices must be
low so the battery life improves, performance improve, reliability improves and
reduces heat removable costs. Because of these reasons the optimization for lower
power dissipation and faster device performance is of prime concern in any design.
The ideal design is one which consumes minimum power, requires minimum area but
has the highest throughput. However, these parameters (area, speed and power) are
often in disagreement and a suitable solution for this is to maintain a tradeoff between
these parameters. At every level of digital design flow there is a possibility in
optimization of power even though benefits are good at the algorithmic and
architectural design level.
In Modern day microprocessors design, maximum speed and low power with
minimum area are main design constraints to meet performance requirement with
technology. An Arithmetic and logic unit (ALU) is a digital electronic circuit that
performs arithmetic and bitwise logical operations on integer binary numbers. ALU
is the heart and crucial building block of all computationally intensive units such as
central processing unit (CPU), Floating point unit (FPU) and Graphical processing
unit (GPU), almost always falls in the data path during the execution of an instruction
decoded by instruction decoder. To meet these requirements, a power efficient, high
performance ALU need to be designed. Hence the power consumption of the ALU
should be kept at a minimum.
2. SOURCES OF POWER DISSIPATION
There are several factors contributing to the CPU power consumption; they include
dynamic power consumption, short-circuit power consumption, and power loss due to
transistor leakage currents.
PCPU = Pdyn + Psc + Pleak
These powers are mainly classified as below two powers such as:
2.1. Dynamic power dissipation
Dynamic power dissipation of CMOS circuit has two parts - dynamic switching
power and short circuit current power [2].
Dynamic switching power is dissipated every time the logic state of the gate
changes. It is represented as P = nfCVdd2, where f is the frequency of switching, CL
is the load capacitance, Vdd is supply voltage and n is the probability of switching.

J. Sneha, Dr.A.S.Srinivasa Rao and A.Rajesh
This power can be reduced by lowering switching frequency; however it is not
desirable as it limits the speed of operation of the device. n can be reduced by
reducing redundant switching activity. Vdd can also be reduced, however it leads to
increased propagation delays and hence not desirable. Hence a proper tradeoff must
be met between these parameters to obtain satisfactory device performance.
Short circuit current power is dissipated when both the NMOS and CMOS
MOSFETs are partially on, during a switching activity. In this case a direct short
circuit path is momentarily formed between power supply and ground, leading to
significant power dissipation. This can be controlled by regulating the slew rate and
applying sharp clock edges. However, generating such a clock is difficult.
2.2. Static or quiescent power dissipation
Static or quiescent power dissipation is independent of the switching activity of the
circuit. This is caused due to leakage current in the device during steady state. Sub-
threshold conduction is the reason for this power dissipation and can be controlled by
biasing the MOSFETs well below their threshold voltages and using multiple
threshold CMOS designs.
In this paper a 16 bit ALU is designed in VERILOG and clock gating (CG)
technique is used to achieve lower power consumption is discussed in section III. And
Also carry select adder with variable block length is used as the primary
computational element of the arithmetic unit for further power reduction is discussed
in section IV. The design is then simulated in ISim simulator and finally implemented
in Xilinx Spartan 3E FPGA. The results are compared with that of a conventional
ALU without clock gating, which revealed significant improved performance of the
clock gated ALU over the conventional design is discussed in section. The results are
shown in section VI. The Concluding remarks and future scope of the technique is
discussed in section VII and VIII.
3. CLOCK GATING
Power can be reduced in several ways such techniques include voltage reduction,
frequency reduction, capacitance reduction, clock gating, using adiabatic logic and
energy recovery logic. The motivation of using the clock gating (CG) technique is: as
the ALU is the unit that performs arithmetic and logical operations, and no both the
operations are performed at the same time. Hence, the clock can be provided to that
unit which needs to be functional. This concept leads to the application of clock
gating technique to the ALU. Another reason, Clock power is a major component of
dynamic power dissipation [3]. In a synchronous circuit several modules are clocked
at the same time. However, at any particular instant only single module may be
functional. Hence the unnecessary clocking of the other modules lead to a lot of
power dissipation. In a clock gating clock is selectively stopped for a portion of
circuit which is not performing any active computation by adding extra hardware to
the circuit. Hence the unnecessary charging and discharging of the unused circuit that
do not perform any active computation is avoided. Thus the clock gating technique is
used to reduce the dynamic power.
 There are many clock gating styles available to optimize power in VLSI circuits.
They can be:
 Latch-free based CG design.
 Latch-based CG design.

3.1. Latch-free Clock Gated ALU design
Clock gating is achieved by ANDing the clock signal with a control signal to form a
gated clock, which is then applied to different components of the circuit. To which
module the gated clock should be applied is decided based on the control signal.
Clock Gate = Clock signal [Gate (either AND or OR)] control signal
Figure 1 Latch-free Clock Gated ALU design
We use an AND gate in clock gate if clock is active on the rising edge. We use an
OR gate in clock gate if clock is active on the falling edge. Using idea given in [4]
and [5], we develop following ALU design as shown in Figure 1.
3.2. Problem in Latch-Free Clock Gated Design
If enable signal goes inactive in between the clock pulse then gated clock terminated
before his life time as shown in Figure 2.
Figure 2 Problem in Latch-Free Clock Gated Design
3.3. Latch Based Clock Gated ALU Design
The latch-based clock gate consists of a level sensitive latch in design to hold the
enable signal from the active edge to the inactive edge of the clock as shown in Figure
3.
Figure 3 Latch Based Clock Gated ALU Design.
This paper implements the latch free CG for a 16-bit ALU on a Spartan 3E FPGA.
The RTL implementation and the power analysis is presented in the paper.

4. LOW-POWER AND AREA-EFFICIENT CARRY SELECT
ADDER
For low power consumption ALU, we need an architecture which has an efficient
adder for propagation and generation block. The sum for each bit position in an
elementary adder is generated sequentially only after the previous bit position has
been summed and a carry propagated into the next position.
Carry Select Adder (CSLA) is one of the fastest adders used in many data-
processing processors to perform fast arithmetic functions. The CSLA is used in many
computational systems to overcome the problem of carry propagation delay by
independently generating multiple carries and then select a carry to generate the sum.
However, the CSLA shown in figure 4 is not area efficient because it uses multiple
pairs of Ripple Carry Adders (RCA) to generate partial sum and carry by considering
carry input cin = 0 and cin = 1, then the final sum and carry are selected by the
multiplexers (mux).
From the regular structure of the CSLA, it is clear that there is scope for reducing
the area and power consumption in the CSLA. This work uses a simple and efficient
gate-level modification to significantly reduce the area and power of the CSLA. The
proposed design has reduced area and power as compared with the regular CSLA with
only a slight increase in the delay.
The basic idea of this proposed structure shown in figure 5 is to use Binary to
Excess-1 Converter (BEC) instead of RCA with cin = 1 in the regular CSLA to
achieve lower area and power consumption [6]-[7]. The main advantage of this BEC
logic comes from the lesser number of logic gates than the n-bit Full Adder (FA)
structure of RCA.
Figure 4 Regular structure of carry select adder.

Figure 5 Proposed structure of carry select adder.V.ALU
Figure 6 Conventional ALU.
We have introduced conventional ALU shown in figure 6 to compare the results
of proposed ALU shown in figure 8. ALU has 5 inputs. They are A, B, sel, reset and
clk. Out returns the result of the ALU operation. A and B are two 16 bit data inputs to
ALU. Input sel [3:0] determines which operation of ALU is to be performed based on
fetched opcode by instruction decoder. The whole process is carried out based on
reset input.
Table 1 Functions of arithmetic and logic unit
Unary sel Arithmetic & logic sel
clear 0000 Adder(carry select adder) 0111
Complement a 0001 Subtraction 1000
Complement b 0010 Add with carry 1001
increment 0011 Subtract with barrow 1010
decrement 0100 Logical and 1011
Left shift 0101 Logical or 1100
Right shift 0110 Logical xor 1101
Logical xnor 1110
Logical nand 1111
As listed in the above table 1, there are 16 functions based on 4 bit opcode. The
opcode selects the particular arithmetic or logical operations. The first 7 are unary
operations and next 9 are arithmetic and logical operations. The ALU generates 3
flags-Zero (Z), Carry (C) and Sign (S) .Z-whose result is zero for any operation
performed on 16 bit ALU, Carry (C) generates a carry when an addition is performed

on 16 bit or generates a borrow when subtraction is performed on 16 bit data, Sign (S)
is generated when the result of any operation is negative.
Figure 7 ALU with clock gating.
Here enable is extra input for generation of clock gate as shown in Figure 7. When
clock is high, then clock and en fed to AND gate and when clock is low, then clock
and en fed to OR gate. The output of AND/OR gate is gated clock. The operation of
proposed model was clearly depicted with the following logic flow chart shown in
figure 8.
Figure 8: Logic flow chart of proposed model.
In brief how the clock gating plays a role in reduction of power is shown with
increment function from conventional to proposed in below figure 9.

Figure 9 Increment function
5. RESULTS
This section shows simulation and implementation results followed by experimental
results of conventional and proposed models. Figure 10 and 11 depicts the timing
waveform and RTL schematic of ALU without clock gating. While figure 12 and 13
depicts the timing waveform and RTL schematic of adder implemented in proposed
model. There by figure 14 and 15 depicts the timing waveform and RTL schematic of
ALU with clock gating.

Figure 10 Simulated waveform of conventional model.
Figure 11 RTL schematic of conventional model.

Figure 12 Simulated waveform of CSLA.
Figure 13 RTL schematic of CSLA.

Figure 14 Simulated waveform of proposed model.
Figure 15 RTL Schematic of proposed model.

Table 2 Device utilization summery of conventional model
Table 3 Device utilization summery of proposed model
Table 2 and 3 shows the implementation results of both the modelsusing
SPARTAN 3E FPGA, realizes that clock gating technique adds the extra hardware to
circuit and reduction in power is shown with table 4.it is formulated as
Table 4 Power comparision of both models
Parameter Conventional model (mW) Proposed model (mW)
Total power 209 198
Dynamic power 130 119
Quiescent power 79 79
Power reduction % = (number of units gated / Total number of units) × 100
Theoretically expected value is (15/16)*100=93.75% power reduction.
Thus by observation on an average the power reduction is 33% for lower
frequency while it is nearly 70% for higher frequencies.
6. CONCLUSION
Finally we have designed above models in VERILOG and simulated with XILINX
ISE 12.1 design suit and then implemented using XILINX SPARTAN 3E FPGA and
results of both the models are compared. From the results 16 bit ALU with clock
gating consumes less power than conventional ALU .Hence it can be concluded that
proposed model uses clock gating is a efficient method for reduction in clock and
dynamic power by reducing unnecessary switching activities on data and clock buses
of unused module while working on current module. This is evidence that how the
Power optimization, traditionally relegated to the synthesis, and placement and
routing stages, has moved up to the System level and RTL stages.
7. FUTURE SCOPE
Clock gating is best technique to reduce dynamic power. In this work, we
implemented our design on 90nm Spartan-3E FPGA. Due to scalability in technology,
Latest FPGA’s like Virtex-7, Kintex-7, Artex-7 based on 28-nm technology
contribute significant leakage power consumption. so there is a scope of using this
technique in 28nm technology to reduce leakage powers. And here target design is for
16-bit ALU, so there is scope to design this ALU in form of 16-bit or 32-bit or even
higher 64-bit.

REFERENCES
[1] Anju S. Pillai, Isha T. B, Factors causing power consumption in an Embedded
processor”, International Journal of Application and Innovation in Engineering
and Management (IJ1IEM), 2(7), July 2013.
[2] KanikaKaur and Arti Noor (2011, May). Strategies and Methodologies for low
power VLSI designs: A review. International Journal of Advances in Engineering
& Technology. [Online]. ISSN: 2231-1963. pp. 159-164. Available:
http://www.ijaet.org/media/18 Strategies-Methodologies-For-Low-Power-Vlsi-
Designs-A-Review-Copyright-IJAET.pdf.
[3] AnkitMitra, Design and implementation of low power 16 bit ALU with clock
gating, International Journal of Advanced Research in Computer Engineering &
Technology (IJARCET), 2(6), June 2013.
[4] J. Shinde and S. S. Salankar, Clock gating-A power optimizing technique for
VLSI circuits, in Proc. Annual IEEE India Conference (INDICON), pp. 1–4,
2011.
[5] V. Khorasani, B. V. Vahdat, and M. Mortazavi, Design and implementation of
floating point ALU on a FPGA processor, IEEE International Conference on
Computing, Electronics and Electrical Technologies (ICCEET), pp. 772–776,
2012.
[6] B. Ramkumar, H.M. Kittur, and P. M. Kannan, ASIC implementation of
modified faster carry save adder, Eur. J. Sci. Res., 42(1), pp. 53–58, 2010.
[7] Y. Kim and L.-S. Kim, 64-bit carry-select adder with reduced area, Electron. Lett,
37(10), pp. 614–615, May 2001.
[8] Bangaru Kalpana Amrut Anilrao Purohit And R. Venkata Siva Reddy, Area
Optimization of SPI Module Using Verilog Hdl, International Journal of
Electronics and Communication Engineering & Technology, 7(3), 2016, pp. 38–
45.
[9] Ms.Kshitija S. Patil ,Prof. G.D.Salunke and Mrs.Bhavana L. Mahajan, FPGA
Implemented Multichannel HDLC Transceiver, International Journal of
Electronics and Communication Engineering & Technology, 3(3), 2012, pp. 170–
176.
[10] Devanshi S. Desai and Dr. Nagendra P. Gajjar, Low Bitrate Modulator Using
FPGA, International Journal of Electronics and Communication Engineering &
Technology, 5(4), 2014, pp. 89–94.

ABOUT THE AUTHORS
Mrs. JAMI. SNEHA is presently pursuing herM.Tech in VLSI
system design in Electronics and Communication Engineering
Department, AITAM, Tekkali. Her areas of interest are Low Power
VLSI system design and Digital Electronics. She has membership in
GSM IEEE.
Dr. ADARI. SATYASRINIVASARAO completed his M. Tech. in
2004 and Ph.D. in 2014, from Andhra University, Vishakapatnam.
He is having 20 years of teaching experience in various engineering
colleges. Presently he is working as Professor, Dept. of ECE,
AITAM, Tekkali. His interested areas, Signal processing, Adaptive
antenna arrays and Communication systems. He has published more
than 25 papers in various journals/proceedings in the field of
Antenna Arrays, Signal processing and Communications.
ATTADA RAJESH completed his M.Tech in 2013 from jawaharlal
nehru technological university, kakinada. He is having 20 years of
teaching experience. Presently he is working as Assistant Professor,
Dept. of ECE, AITAM, Tekkali. His interested areas, DSP, VLSI,
Image Processing, Antennas, E M waves, Radar systems. He has
published 5 papers in various journals/proceedings in the field of
signal and image processing.

DESIGN AND IMPLEMENTATION OF LOW POWER ALU USING CLOCK GATING AND CARRY SELECT ADDER

Recommended

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to DESIGN AND IMPLEMENTATION OF LOW POWER ALU USING CLOCK GATING AND CARRY SELECT ADDER (20)

More from IAEME Publication (20)

Recently uploaded (20)

DESIGN AND IMPLEMENTATION OF LOW POWER ALU USING CLOCK GATING AND CARRY SELECT ADDER