Optimized Design of an Alu Block Using Power Gating Technique

S. Anvesh Int. Journal of Engineering Research and Applications www.ijera.com
ISSN: 2248-9622, Vol. 5, Issue 8, (Part - 5) August 2015, pp.204-209
www.ijera.com 204 | P a g e
Optimized Design of an Alu Block Using Power Gating Technique
S. Anvesh1
and P. Ramana Reddy2
1
M. Tech student and 2
Associate Professor & Head Department of Electronics & Communication Engineering
Jawaharlal Nehru Technological University Anantapur, India
ABSTRACT
Power is the limiting factor in traditional CMOS scaling and must be dealt with aggressively. With the scaling
of technology and the need for high performance and more functionality, power dissipation becomes a major
bottleneck for a system design. Power gating of functional units has been proved to be an effective technique to
reduce power consumption. This paper describe about to design of an ALU block with sleep mode to reduce the
power consumption of the circuit. Local sleep transistors are used to achieve sleep mode. During sleep mode
one functional unit is working and another functional unit is in idle state. i.e., it disconnects the idle logic
blocks from the power supply. Architecture and functionality of the ALU implemented on FPGA and is tested
using DSCH tool. Power analysis is carried out using MICROWIND tool.
Keywords: Low Power, Power gating technique, Sleep mode, FPGA
I. INTRODUCTION
The increasing speed and complexity of today’s
designs implies a significant increase in the power
consumption of very-large-scale integration circuits.
Voltage is increased to attain higher speed of
operation with the penalty of increased power
consumption. To meet this challenge, researchers
have developed many different design techniques to
reduce power. CMOS technology scaling has now
moved to a power constrained condition. Circuit
techniques to reduce chip standby leakage have
become a key enabler. Scaling is a trade off between
performance and leakage. One of the key features
that led to the success of complementary metal-oxide
semiconductor technology was its intrinsic low-
power consumption [2]. This meant that circuit
designers and electronic design automation (EDA)
tools could afford to concentrate on maximizing
circuit performance and minimizing circuit area.
Another interesting feature of CMOS technology is
its nice scaling properties, which has permitted a
steady decrease in the feature size, allowing for more
and more complex systems on a single chip, working
at higher clock frequencies [1]. In fact, power
consumption is regarded as the limiting factor in the
continuing scaling of CMOS technology.
II. NEED FOR LOW POWER
As the demand for high speed, Low power
consumption and high performance continues to
grow each year, there is a need to scale the devices to
smaller dimensions. As the market trend moves
towards a greater scale of integration, the move
toward a reduction in supply voltage also have the
advantage of improving the reliability of the chip. As
the complexity of the chip increases design problems
occurs such as layout of a chip and simulation for the
circuit and these are all becomes more intense.
Power consumption is regarded as the limiting factor
in the present technology. High power supply affects
the temperature and further influences the cooling
cost of the chip. If the temperature increases it
directly reflects the environmental problems. Noise
and reliability [3] is the one more big issue in the
field of high power consumption devices. With all
this, high power consumption will defiantly degrade
Battery life of the electronics devices.
III. Low-Power Design Methodologies
A significant amount of recent research has
sought to reduce the power consumption. Power
consumption can be reduced at many levels, from the
circuit level to the application level. This paper seeks
to explore the limits of power reduction available due
to optimizations at the architecture level. This work
identifies three broad categories of wasted power that
either has been potentially be targeted for reduction.
Those categories are process technology, reducing
switching activity and power down modes. The focus
of this research is reduction of power dissipation in
Architectural level. These categories of research in
architectural power reduction assuming solutions that
do not fundamentally hang the system or its
performance [3], [4].
A. Process Technology
The adaptation of process technology will place
a major role in low level system design. It includes
the technologies of reduction of capacitance and
reducing leakage power, supply power and higher
density of integration.
B. Reducing Switching Activity
RESEARCH ARTICLE OPEN ACCESS

To reduce power dissipation effectively low
power methodologies can be used. Since designers
have no influence on Vdd, and switching activity is
the remaining one to influence. Methodologies for
reducing the switching activity are minimization of
glitches and minimization of number of operations,
using low power bus. Scaling down features requires
lowering the threshold voltage, which increases
leakage power.
C. Power down Modes
When some modules of the program are idle and
do not add any functionality to the design at the time
but still consume power, that system typically is not
working at maximum possible performance. The
power down modes will affect the design architecture
mode and it plays a major role in the power
dissipation of the circuit. Mostly use the power down
modes like power gating and clock gating.
IV. Low Power Techniques For Field
Programmable Gate Arrays
A) Introduction
Field-programmable gate arrays are ideal for
adaptive systems, since they are reconfigurable and
can be programmed to implement any digital logic.
Applications of such FPGA based adaptive systems
include face image recognition, on-line failure
recovery and analysis of fire fly synchronization.
Traditionally, FPGA research focused on reducing
the speed and area overhead. In recent years, much of
the focus has shifted to improving the energy
efficiency [9]. This shift is due to process scaling and
increased demand for low power applications.
Although process scaling reduces the energy needed
to perform a given computation since wires and
transistors are smaller, it increases power dissipation
per unit area and therefore the overall power for a
given die size
B) Power dissipation in CMOS
In digital circuits power dissipation is
categorized into two types: peak power and time-
averaged power consumption. Peak power is a
reliability issue that determines both the chip lifetime
and performance.[6]The voltage drop effects and
Transistor sizing will affect the performance of a
design due to the increased gate and interconnect
delay. This large power consumption causes the
device to overheat which reduces the reliability and
lifetime of the circuit. The time-averaged power
consumption in conventional CMOS digital circuits
occurs in two forms: dynamic and static [4] .
a) Dynamic power dissipation Dynamic power
dissipation occurs in the logic gates that are in
the process of switching from one state to
another. During this process, any internal and
external capacitance associated with the gate's
transistors has to be charged, thereby consuming
power.
b) Static power dissipation Static power
dissipation is associated with inactive logic gates
i.e. not currently switching from one state to
another. Static power is caused by leakage
currents while the gates are idle that is, no output
transitions. Static power is important during
standby especially for battery-powered devices.
C) Basic Power down Mode Techniques
a) Clock Gating
Clock power is an important component of
overall dynamic power. One way to reduce clock
power is to use clock gating, which dynamically
disables the clock in unused parts of the circuit. This
avoids the unnecessary power dissipation caused by
charging and discharging the clock signal at these
unused gates [3].Gating is generally achieved by
ANDing the clock signal with a clock-gate signal,
which rises whenever the part of the circuit that needs
to be gated is active and remains low otherwise.
b) Power Gating
Power Gating has become one of the most
widely used circuit design techniques for reducing
leakage current in Static power dissipation. [3] This
uses high threshold-voltage transistors or sleep
transistors in circuit blocks that switch in frequently.
Its concept is very simple, but its application to
standard-cell VLSI designs involves many careful
considerations. The great complexity of designing a
power-gated circuit originates from the side effects of
inserting current switches, which have to be resolved
by a combination of extra circuitry and customized
tools and methodologies.
V. Power Gating Methodology
Power-gating technique uses a high Vth sleep
transistor in series with the pull-up and/or the pull-
down of a low Vth logic block to reduce leakage
power. The sleep transistor can be turned off when
the low Vth logic block is inactive, thus resulting in a
significant reduction of sub-threshold leakage
current. Figure 1 shows a sleep transistor used for
power gating. A sleep transistor can be a high Vth
NMOS or PMOS transistor [6]. A PMOS sleep
transistor served as a header switch connects the
power network to virtual VDD. An NMOS sleep
transistor served as a footer switch connects the
ground network to virtual GND. Normally, either a
header switch or a footer switch is used to conserve
area and reduce timing penalty caused by voltage
drop across sleep transistors.
In this paper, consider only footer switch. The
logic block in Figure 1 can be as simple as a logic
gate. If it is a logic gate in a standard cell library, we
call it a power gated cell. If the power gated cells in a

standard cell design are used in an isolated manner,
i.e., their virtual VDDs are not connected; call this
power gating approach fine-grained power gating [3].
If a sleep transistor is shared by a cluster of power-
gated logic cells, possibly located in close proximity,
such kind of power gating is called coarse-grained or
cluster-based power gating. If the virtual VDDs of
power gated cells (clusters) in a design are connected
together, the sleep transistors of these power gated
cells (clusters) are in fact form a transistor network.
We call such kind of power gating distributed sleep
transistor network.
Figure 1. Power gating structure
 Power Gating Considerations:
 Library design: Special cells are needed like
switches, isolation cells Power Gating
Considerations.
 Headers or Footers: What type of switch have
to used either small or big and what type of
connection use either header or footer
connection.
 Which modules, and how many, to be power
gated.
 Sleep control signal must be available, or must
be created.
 Switches affect source voltages which affect
delays
 Power-up & power-down sequencing.
 Controller must be designed and sequencing
verified.
 Transistor Size: The size of the transistor is
measured in terms of W/L ratio. This is also
called aspect ratio of transistor. While designing
a Power gate, the size of the sleep transistor or
switch is also one important consideration.
In the function mode, a sleep transistor is
operated in the linear region and can be modeled as
resistance and there will be a voltage drop across
sleep transistor. Voltage drop can have performance
impact and normally, the voltage needs to be
confined to certain percentage of VDD such as 5 to
10% of VDD [5].
VI. Alu Design Through Power Gating
Methodology
The ALU has four stages, each stage consisting
of three parts: a) input b) internal modules and c)
output multiplexer. The ALU performs the following
four arithmetic operations, ADD, OR, FULL
ADDER, INVERTER. The output sections consist of
4 to 1 and 8 to 1 and 16 to 1 multiplexers. These
multiplexers were designed using the 2 to 1
multiplexer logic. A set of two select signals has been
incorporated in the design to determine the operation
being performed and the inputs and outputs being
selected.
As from the technique of power gating the
proposed ALU Design is shown in Figure 2. ALU
logic block is connected to the sleep transistor
network. This sleep transistor network functionality
is based on the sleep signal logic.
Figure 2. Modified ALU Architecture
Traditional 1 bit ALU:
Figure 3. Traditional 1 bit ALU Architecture
As in the power analysis we first design the
circuit and after that power estimation is done. Here
the estimated power consumed is 2.997µW.

Optimized 1 bit ALU:
Upgraded ALU architecture is shown in figure 4.
As in the case of Optimized ALU, sleep transistor
mode of switching technique is being used.
NMOS is used as a sleep transistor and this
NMOS is connected to the S1. When ever the S1
turns to ON state then the supply rail will be
connected to the NMOS transistor and our Cin will be
connected to full adder block.
Figure 4. Optimized 1 bit ALU Architecture.
The same procedure is applied to the Optimized
ALU also, i.e. design the Optimized ALU circuitry
and then estimation of power is done. This estimation
of power is 1.277µW.
For traditional 4 bit ALU design, FOUR 1 bit
ALU’s are cascaded like as shown in figure 5. For
this circuitry also same procedure will be apply.
The estimated power for the Traditional 4 bit
ALU is20.24µW. Optimized 4 bit ALU design is
shown in figure6. There is a significant reduction in
power is observed. The power estimated
was5.467µW.
Figure 6. Optimized 4 bit ALU Architecture
Traditional 8 bit ALU architecture is shown in
Figure 7. Here EIGHT 1 bit Traditional ALUs are
cascaded to perform 8 bit operation. The estimated
Power is 30.99µW.
For Optimized 8 bit ALU Architecture cascading
of EIGHT 1 bit Optimized ALUs is done. The design
of this Architecture is shown in Figure 8. The
estimated power for this Optimized 8 bit ALU is
9.874µW.
Figure 8. Optimized 8 bit ALU Architecture.

VII. SIMULATION RESULTS
Figure 9. Waveform for Traditional 1 bit ALU
Figure 10. Waveform for Optimized 1 bit ALU
Figure 15. Power dissipation comparisons of
traditional and optimized methods
VIII. CONCLUSION
This paper reviewed few key circuit design
techniques and the corresponding implications on
Power optimization. At first, designed the Traditional
architecture of ALU and then optimized the
Architecture using power gating technique to reduce
the Power level. From this techique there is a
significant reduction in power is observed. And it
was a possible thing that the extenssion of this paper
for 32 bit and 64 bit also.
REFERENCES
[1.] J.Adam Butts and Gurindar S. Sohi. “A
Static Power model for Architects”. In
MICRO – 33 Proceedings of the 33rd
Annual IEEE/ACM International

Symposium on Micro architecture, 2000. pp.
191-201.
[2.] M.Chang,C.Chang,C.Chao,K.Goto,M.Leong
,L.Lu,and C.Diaz, “Transistor- and Circuit-
Design Optimization for Low-Power
CMOS”, IEEE Transactions on Electron
Devices, January 2008.
[3.] Richa Srivastava, S.A .Imam, Sujata
Pandey, “Low Power Design Techniques for
high performance Digital Integrated
Circuits,” MASAUM Journal of Reviews
and Surveys,Vol. 1, No.1, 2009.
[4.] Frank Propen, “Low power design guide”
reference book.
[5.] Neil H.E.,David Harris, “CMOS VLSI
Design” Third Edition, Boston: Pearson,
2005 .
[6.] Anantha P. Chandrakasan and Robert W.
Brodersen “Minimizing Power Consumption
in CMOS Circuits” Department of EECS,
University of California at Berkeley.
[7.] B. Davari, R. H. Dennard and G. G. Shahidi,
“CMOS Scaling for High Performance and
Low Powe The Next Ten Years,” Proc.
IEEE, April 1995, pp. 595-606.
[8.] Power Optimization [Online]. Available:
http://en.wikipedia.org/wiki/Power_optimiza
tion_(EDA).
[9.] Power optimizations in FPGA designs
[online]http://www.altera.comcp-
pwropt.pdf.

Optimized Design of an Alu Block Using Power Gating Technique

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Optimized Design of an Alu Block Using Power Gating Technique (20)

Recently uploaded (20)

Optimized Design of an Alu Block Using Power Gating Technique