www.vidyarthiplus.

com

Einstein College of
Engineering

EC64 VLSI DESIGN SYLLABUS

UNIT I CMOS TECHNOLOGY
A brief History-MOS transistor, Ideal I-V characteristics, C-V characteristics, Non ideal IV
effects, DC transfer characteristics - CMOS technologies, Layout design Rules, CMOS
process enhancements, Technology related CAD issues, Manufacturing issues

UNIT II CIRCUIT CHARACTERIZATION AND SIMULATION

Delay estimation, Logical effort and Transistor sizing, Power dissipation, Interconnect,
Design margin, Reliability, Scaling- SPICE tutorial, Device models, Device characterization,
Circuit characterization, Interconnect simulation
UNIT III COMBINATIONAL AND SEQUENTIAL CIRCUIT DESIGN
Circuit families Low power logic design comparison of circuit families Sequencing static
circuits, circuit design of latches and flip flops, Static sequencing element methodologysequencing dynamic circuits synchronizers
UNIT IV CMOS TESTING
Need for testing- Testers, Text fixtures and test programs- Logic verification- Silicon
debug principles- Manufacturing test Design for testability Boundary scan
UNIT V SPECIFICATION USING VERILOG HDL
Basic

concepts-

identifiers-

gate

primitives,

gate

delays,

operators,

timing

controls,

procedural assignments conditional statements, Data flow and RTL, structural gate level switch
level modeling, Design hierarchies, Behavioral and RTL modeling, Test benches, Structural gate
level description of decoder, equality detector, comparator, priority encoder, half adder, full adder,
Ripple carry adder, D latch and D flip flop.
Textbooks:
1. Weste and Harris: CMOS VLSI DESIGN (Third edition) Pearson Education, 2005
2. Uyemura J.P: Introduction to VLSI circuits and systems, Wiley 2002.
3. J.Bhasker,Verilog HDl Primer , BS publication,2001(UNIT V)
References:
1 D.A Pucknell & K.Eshraghian Basic VLSI Design, Third edition, PHI, 2003
2 Wayne Wolf, Modern VLSI design, Pearson Education, 2003
3 M.J.S.Smith: Application specific integrated circuits, Pearson Education, 1997
4 Ciletti Advanced Digital Design with the Verilog HDL, Prentice Hall of India, 2003

www.vidyarthiplus.com

UNIT I CMOS TECHNOLOGY

INTRODUCTION
An MOS (Metal-Oxide-Silicon) structure is created by superimposing several layers of
conducting, insulating, and transistor forming materials. After a series of processing steps, a
typical structure might consists of levels called diffusion, polysilicon, and metal that are
separated by insulating layers. CMOS technology provides two types of transistors, an n-type
transistor (n MOS) and a p-type transistor (p MOS). These are fabricated in silicon by using
either negatively doped silicon that is rich in electrons (negatively charged) or positively doped
silicon that is rich in holes (the dual of electrons and positively charged). For the n-transistor, the
structure consists of a section of p-type silicon separating two diffused areas of n-type silicon. The
area separating the n regions is capped with a sandwich consisting of an insulator and a
conducting electrode called the GATE. Similarly, for the p-transistor the structure consists of a
section of n-type silicon separating two p-type diffused areas. The p-transistor also has a gate
electrode. The gate is a control input and it affects the flow of electrical current between the drain
and source. The drain and source may be viewed as two switched terminals.
An MOS transistor is termed a majority-carrier device, in which the current in a
conducting channel between the source and drain is modulated by a voltage applied to the gate. In an
n-type MOS transistor (i.e.,nMOS), the majority carriers are electrons. A positive voltage applied on
the gate with respect to the substrate enhances the number of electrons in the channel (region
immediately under the gate) and hence increases the conductivity of the channel. The operation
of a p-type transistor is analogous to the nMOS transistor, with the exception that the majority
carriers are holes and the voltages are negative with respect to the substrate. The switching
behavior of an MOS device is characterized by threshold voltage, Vt. This is defined as the voltage
at which an MOS device begins to conduct. For gate voltage less than a threshold value, the channel
is cut-off, thus causing a very low drain- to-source current. Those devices that are normally cut-off
(i.e., non-conducting) with zero gate bias are further classed as enhancement mode devices, whereas
those devices that conduct with zero gate bias are called depletion mode devices.

CV Characteristics:
The measured MOS capacitance (called gate capacitance) varies with the applied gate voltage

A very powerful diagnostic tool for identifying any deviations from the ideal in both
oxide and semiconductor

Routinely monitored during MMOS device fabrication

Measurement of C-V characteristics

Apply any dc bias, and superimpose a small (15 mV) ac signal

Generally measured at 1 MHz (high frequency) or at variable frequencies between

1KHz to 1 MHz

The dc bias VG is slowly varied to get quasi-continuous C-V characteristics

DC Characteristics of CMOS inverter

The general arrangement and characteristics are illustrated in Fig. 2.3. The current/voltage
relationships for the MOS transistor may be written as,

Ids = K W

(Vds Vt)Vds Vds

Figure 2.3 CMOS inverter

In the resistive region, or

Ids = K W (Vgs Vt)

L

In the saturation region. In both cases the factor K is a technology- dependent parameter such that
K = ins o
D
The factor W/L is contributed by the geometry and it is common practice to write
=KW
L
Such that,
Ids = (Vgs Vt)

2
In saturation, and where may be applied to both nMOS and pMOS transistors as
follows, n = ins o n
D
p = ins o p
D

Wn

Ln
Wp
Lp

Where Wn and Ln, Wp and Lp are the n- and p- transistor dimensions respectively. The CMOS
inverter has five regions of operation is shown in Fig. 2.4 and in Fig. 2.5.

Figure 2.4 Transfer characteristics

Considering the static condition first, in region 1 for which Vin = logic 0, the p-transistor fully
turned on while the n-transistor is fully turned off. Thus no current flows through the inverter and the
output is directly connected to VDD through the p-transistor.
In region 5 Vin = logic 1, the n-transistor is fully on while the p-transistor is fully off. Again,
no current flows and a good logic 0 appears at the output.
In region 2 the input voltage has increased to a level which just exceeds the threshold voltage
of the n-transistor. The n-transistor conducts and has a large voltage between source and drain. The ptransistor also conducting but with only a small voltage across it, it
unsaturated resistive region.

Figure 2.5 CMOS inverter current versus Vin

operates in the

In region 4 is similar to region 2 but with the roles of the p- and n- transistors reversed. The
current magnitudes in region 2 and 4 are small and most of the energy consumed in switching from
one state to the other is due to the large current which flows in region 3.
In region 3 is the region in which the inverter exhibits gain and in which both transistors are
in saturation.
The currents in each device must be the same since the transistors are in series. So we may
write
I dsp = - Idsn
Where
Idsp= n (Vin VDD - Vtp )

2
And
Idsn = n (Vin Vtn )

2
Vin in terms of the ratio and the other circuit voltages and currents
Vin = VDD + Vtp +Vtn (n + p)

1/2

1/2

1+ (n + p)

Since both transistors are in saturation, they act as current sources so that the equivalent circuit in
this region is two current sources so that the equivalent circuit in this region is two current
sources in series between VDD and VSS with the output voltage coming from their common point.
The region is inherently unstable in consequence and the change over from one logic level to the
other is rapid.
If n= p and if Vin = -Vtp, then
Vin = 0.5 VDD
Since only at this point will the two factors be equal. But for n= p the device geometries must be
such that
p Wp/Lp = n Wn/Ln
The mobilities are inherently unequal and thus it is necessary for the width to length ratio of the
p- device to be three times that of the n-device, namely
Wp/Lp = 2.5 Wn/Ln
The mobility is affected by the transverse electric field in the channel and is thus independent on
Vgs. It has been shown empirically that the actual mobility is
= z (1 (Vgs Vt)

-1

is a constant approximately equal to 0.05 Vt includes anybody effect, and z is the mobility with
zero transverse field.

CMOS Technologies

CMOS TECHNOLOGIES
CMOS provides an inherently low power static circuit technology that has the capability of providing
a lower-delay product than comparable design-rule nMOS or pMOS technologies. The four dominant
CMOS technologies are:
P-well process
n-well process
twin-tub process
Silicon on chip process

The p-well process

A common approach to p-well CMOS fabrication is to start with moderately doped n-type
substrate (wafer), create the p-type well for the n-channel devices, and build the p-channel transistor
in the native n-substrate. The processing steps are,
1. The first mask defines the p-well (p-tub) n-channel transistors (Fig. 1.5a) will be fabricated
in this well. Field oxide (FOX) is etched away to allow a deep diffusion.
2. The next mask is called the thin oxide or thinox mask (Fig. 1.5b), as it defines where
areas of thin oxide are needed to implement transistor gates and allow implantation to form
p- or n- type diffusions for transistor source/drain regions. The field oxide areas are etched
to the silicon surface and then the thin oxide areas is grown on these areas. O ther terms for
this mask include active area, island, and mesa.
3.

Polysilicon gate definition is then completed. This involves covering the surface
with polysilicon (Fig 1.5c) and then etching the required pattern (in this case an inverted
U). Poly gate regions lead to self-aligned source-drain regions.
+

4. A p-plus (p ) mask is then used to indicate those thin-oxide areas (and polysilicon) that are to
+

be implanted p . Hence a thin-oxide area exposed by the p-plus mask (Fig. 1.5d) will become
+

a p diffusion area. If the p-plus area is in the n-substrate, then a p-channel transistor or ptype wire may be constructed. If the p-plus area is in the p-well, then an ohmic contact to the
p-well may be constructed.
5. The next step usually uses the complement of the p-plus mask, although an extra mask is
normally not needed. The absence of a p-plus region over a thin-oxide area indicates
that the area will be an n
possible n-

diffusion or n-thinox. n-thinox in the p-well defines

transistors and wires. An n diffusion (Fig. 1.5e) in the n-substrate allows an ohmic contact to
be made. Following this step, the surface of the chip is covered with a layer of Sio2.
6. Contacts cuts are then defined. This involves etching any Sio2 down to the contacted surface,
these allow metal (Fig. 1.5f) to contact diffusion regions or polysilicon regions.
7. Metallization (Fig. 1.5g) is then applied to the surface and selectively etched.
8. As a final step, the wafer is passivated and openings to the bond pads are etched to allow for
wire bonding. Passivation protects the silicon surface against the ingress of contaminants.

(a)

(b)

(c)

(d)

(e)

(f)

(g)
Figure 1.5 Typical p-well CMOS process steps with corresponding masks required

Basically the structure consists of an n-type substrate in which p-devices may be formed by
suitable masking and diffusion and, in order to accommodate n-type devices, a deep p-well is diffused
into the n-type substrate. This diffusion must be carried out with special care since the p-well doping
concentration and depth will affect the threshold voltages as well as the breakdown voltages of the ntransistors. To achieve low threshold voltage (0.6 to 1.0 V), deep well diffusion or high well
resistivity is needed. However, deep wells require larger spacing between the n- and p-type
transistors and wires because of lateral diffusion resulting in larger chip areas.
High resistivity can accentuate latch-up problems. In order to achieve narrow threshold voltage
tolerances in a typical p-well process, the well concentration is made about one order of
magnitude higher than the substrate doping density, thereby causing the body effect for n-channel
devices to be higher than for p-channel transistors. In addition, due to this higher concentration, ntransistors suffer from excessive source/drain to p-well capacitance will tends to be slower in
performance. The well must be grounded in such a way as to minimize any voltage drop due to
injected cur rent in substrate that is collected by the p-well.
The p-well act as substrate for then-devices within the parent n-substrate, and, provided polarity
restrictions are observed, the two areas are electrically isolated such that there are in affect two
substrate, two substrate connections (VDD and VSS) are required.
The n-well process:
The p-well processes have been one of the most commonly available forms of CMOS. However, an
advantage of the n-well process is that it can be fabricated on the same process line as conventional n
MOS. n well CMOS circuits are also superior to p-well because of the lower substrate bias effects on
transistor threshold voltage and inherently lower parasitic capacitances associated with source and
drain regions.
Typically n-well fabrication steps are similar to a p-well process, except that an n-well is
used which is illustrated in Fig. 1.6. The first masking step defines the n-well regions. This followed
by a low phosphorus implant driven in by a high temperature diffusion step t o form the n-wells.
+

The well depth is optimized to ensure against p-substrate to p diffusion breakdown without
+

compromising the n-well to n mask separation. The next steps are to define the devices and
diffusion paths, grow field oxide,
diffusions,

make

contact

cuts

deposit

and

pattern

the

polysilicon,

carry

out

the

and metallization. An n-well mask is used to define n-well

+

regions, as opposed to a p-well mask in a p- well process. An n-plus (n ) mask may be used to
define the n-channel transistors and VDD contacts. Alternatively, we could use a p-plus mask to
define the p-channel transistors, as the masks usually are the complement of each other.

Figure 1.6 Main steps in a typical n-well process

Due to differences in mobility of charge carriers the n-well process creates non-optimum
p- channel characteristics, such as high junction capacitance and high body effect. The n-well
technology has a distinct advantage of providing optimum device characteristics. Thus n-channel
devices may be used to form logic elements to provide speed and density, while p-transistors could
primarily serve as pull-up devices.
The twin-tub process:
Twin-tub CMOS technology provides the basis for separate optimization of the p-type and n-type
transistors, thus making it possible for threshold voltage, body effect, and the gain associated with
+

n- and p-devices to be independently optimized. Generally the starting material is either an n or

+

p substrate with a lightly doped epitaxial or epi layer, which is used for protection against latch-up.
The aim of epitaxy is to grow high purity silicon layers of controlled thickness with accurately
determined dopant concentrations distributed homogeneously throughout the layer. The electrical
properties for this layer are determined by the dopant and its concentration in the silicon.
The process sequence, which is similar to the p-well process apart from the tub formation
where both p-well and n-well are utilized as in Fig. 1.7, entails the following steps:
Tub formation

 Thin oxide etching

Source and drain implantations
Contact cut definition
Metallization.

Figure 1.7 Flow diagram of twin-tub process

Since this

process

provides separately optimized

wells,

better

performance

n-

transistors (lower capacitance, less body effect) may be constructed when compared with a
conventional p-well process. Similarly the p-transistors may be optimized. The use of threshold
adjust steps is included in this process.
Silicon on insulator process:
Silicon on insulator (SOI) CMOS processes has several potential advantages such as higher
density, no latch-up problems, and lower parasitic capacitances. In the SOI process a thin layer
of single crystal silicon film is epitaxial grown on an insulator such as sapphire or magnesium
aluminate spinel. The steps involves are:
1) A thin film (7-8 m) of very lightly doped n-type Si is grown over an insulator (Fig. 1.8a).
Sapphire is a commonly used insulator.
2) An anisotropic etch is used to etch away the Si (Fig. 1.8b) except where a diffusion area will
be needed.
3)

The p-islands are formed next by masking the n-islands with a photoresist. A p-type
dopant (boron) is then implanted. It is masked by the photoresist and at the u nmasked
islands. The p- islands (Fig. 1.8c) will become the n-channel devices.

4) The p-islands are then covered with a photoresist and an n-type dopant, phosphorus, is
implanted to form the n-islands (Fig. 1.8d). The n-islands will become the p-channel devices.
5) A thin gate oxide (500-600) is grown over all of the Si structures (Fig. 1.8e). This is
normally done by thermal oxidation.
6) A polysilicon film is deposited over the oxide.
7) The polysilicon is then patterned by photomasking and is etched. This defines the polysilicon
layer in the structure as in Fig. 1.8f.
8) The next step is to form the n-doped source and drain of the n-channel devices in the pislands. The n-island is covered with a photoresist and an n-type dopant (phosphorus) is
implanted (Fig. 1.8g).
9) The p-channel devices are formed next by masking the p-islands and implanting a p-type
dopant. The polysilicon over the gate of the n-islands will block the dopant from the gate,
thus forming the p-channel devices is shown in Fig. 1.8h.

10) A layer of phosphorus glass is deposited over the entire structure. The glass is etched at
contact cut locations. The metallization layer is formed. A final passivation layer of a
phosphorus glass is deposited and etched over bonding pad locations.

(a)

(b)

(c)

(d)

(e)
(
f
)

(
g
)

(h)
Figure 1.8 SOI fabrication steps
The advantages of SOI technology are:
Due to the absence of wells, denser structures than bulk silicon can be obtained.
Low capacitances provide the basis of very fast circuits.
No field-inversion problems exist.
No latch-up due to isolation of n- and p- transistors by insulating substrate.
As there is no conducting substrate, there are no body effect problems
Enhanced radiation tolerance.
But the drawback is due to absence of substrate diodes, the inputs are difficult to protect. As device
gains are lower, I/O structures have to be larger. Single crystal sapphires are more expensive
than silicon and processing techniques tend to be less developed than bulk silicon techniques.
BiCMOS TECHNOLOGY FABRICATION
The MOS technology lies in the limited load driving capabilities of MOS transistors. This is due to
the limited current sourcing and current sinking abilities associated with both p- and n- transistors.
Bipolar transistors provide higher gain and have generally better noise and high frequency
characteristics than MOS transistors and have effective way of speeding up VLSI circuits. When
considering CMOS technology, there is difficulty in extending the fabrication processes to include
bipolar as well as MOS transistors. Indeed, a problem of p-well and n-well CMOS processing is that
parasitic bipolar transistors are formed as part of the outcome of fabrication.

The production of npn bipolar transistors with good performa nce characteristics can
be achieved by extending the standard n-well CMOS processing to include further masks to add
+

two additional layers such as the n subcollector and p base layers. The npn transistors is formed in
+

an n- well and the additional p base region is located in the well to form the p-base region of the
+

transistor. The second additional layer, the buried n subcollector (BCCD), is added to reduce
the n-well (collector) resistance and thus improve the quality of the bipolar transistor. The
arrangement of BiCMOS npn transistor is shown in Fig. 1.9.

Figure 1.9 Arrangement of BiCMOS npn transistor

There are several advantages if the properties of CMOS and bipolar technologies
could be combined. This is achieved to a significant extent in the BiCMOS technology. A further
advantage which arises from BiCMOS technology is that analog amplifier design is facilitated
and improved. High impedance CMOS transistors may be used for the input circuitry while the
remaining stages and output drivers are realized using bipolar transistors. Since extra design and
processing steps are involved as in Fig. 1.10, there is an increase in cost and some loss of packing
density.

Figure 1.10 n-well BiCMOS fabrication steps

UNIT II CIRCUIT CHARACTERIZATION AND SIMULATION

Delay estimation:
Estimation of the delay of a Boolean function from its functional description is an important step
towards design exploration at the register transfer level (RTL). This paper addresses the problem
of estimating the delay of certain optimal multi-level implementations of combinational circuits,
given only their functional description. The proposed delay model uses a new complexity measure
called the delay measure to estimate the delay. It has an advantage that it can be used to predict
both, the minimum delay (associated with an optimum delay implementation) and the
maximum delay (associated with an optimum area implementation) of a Boolean function without
actually resorting to logic synthesis. The model is empirical and results demonstrating its
feasibility and utility are presented.
tpdr: rising propagation delay
From input to rising output crossing VDD/2
tpdf: falling propagation delay
From input to falling output crossing VDD/2
tpd: average propagation delay
tpd = (tpdr + tpdf)/2
tr: rise time
From output crossing 20% to 80% VDD
tf: fall time
From output crossing 80% to 20% VDD
tcd: average contamination delay
tcd = (tcdr + tcdf)/2
tcdr: rising contamination delay: Min from input to rising output crossing VDD/2
tcdf: falling contamination delay: Min from input to falling output crossinVDD/2
Solving differential equations by hand is hard. SPICE like simulators used for accurate analysis.
But simulations are expensive. We need to be able to estimate delay although not as accurately as
simulator.
Use RC delay models to estimate delay
C = total capacitance on the output node
Use Effective resistance R
Therefore tpd = RC
Transistors are characterized by finding their effective R.

Transistor sizing:

Not all gates need to have the same delay.

Not all inputs to a gate need to have the same delay.

Adjust transistor sizes to achieve desired delay.

Example: Adder carry chain

Inter-stage effects in transistor sizing

Increasing a gates drive also increases the load to the previous stage
Logical effort
Logical effort is a gate delay model that takes transistor sizes into account. Allows us to optimize
transistor sizes over combinational networks. Isnt as accurate for circuits with reconvergent fanout.
Logical effort gate delay model
Express delays in process-independent unit
Gate delay is measured in units of minimum-size

inverter delay .
Gate delay formula:

d = f + p.
Effort delay f is related to gates load. Parasitic delay p depends on gates structure. Represents
delay of gate driving no load Set by internal parasitic capacitance
Effort delay
Effort delay has two components: f = gh.
Electrical effort h is determined by gates load: h = Cout/Cin Sometimes called fanout
Logical effort g is determined by gates structure. Measures relative ability of gate to deliver
current g 1 for inverter
Delay plots:

Computing Logical Effort

Logical effort is the ratio of the input capacitance of a gate to the input capacitance of an
inverter delivering the same output current. Measure from delay vs. fanout plots Or estimate by
counting transistor widths

Power Estimation:
The past the major concerns of the VLSI designer were area performance cost and
reliability power considerations were mostly of only secondary importance. In recent years
however this has begun to change and increasingly power is being given comparable weight to area
and speed. Several factors have contributed to this trend Perhaps the primary driving factor has
been the remarkable success and growth of the class of personal computing devices portable
desktops

audio

and videobased multimedia products_ and wireless communications systems

_personal digital assistants and personal communicators_ which demand high_speed computation
and complex functionality with In the past_ the major concerns of the VLSI designer were area_
performance cost and reliability_ power considerations were mostly of only secondary importance_
In recent years_ however_ this has begun to change and_ increasingly_ power is being given
comparable weight to area and speed_ Several factors have contributed to this trend Perhaps
the primary driving factor has been the remarkable success and growth of the class of personal
computing devices _portable desktops_ audio_ and

video_based

multimedia

products

and

wireless communications systems _personal digital assistants and personal communicators

which

demand

high_speed

computation

and

complex functionality with low power

consumption_ There also exists a strong pressure for producers of high_end products to reduce
their power consumption.
Software_Level Power Estimation
The first task in the estimation of power consumption of a digital system is to identify the
typical application programs that will be executed on the system. A non_trivial application program
consumes millions of machine cycles_ making it nearly impossible to perform power estimation
using the complete program at_ say_ the RT_level_ Most of the reported results are based on
power macro_modeling_ an estimation approach which is extensively used for behavioral and
RTL level estimation see Sections and In the power cost of a CPU module is characterized by
estimating the average capacitance that would switch when the given CPU module is
activated_In the switching activities on _address_ instruction_ and data_ buses are used to estimate
the power consumption of the microprocessor, based on actual current measurements of some
processors_ Tiwari et al_ present the following instruction_level power model

where Energyp is the total energy dissipation of the program which is divided into three parts The
first part is the summation of the base energy cost of each instruction, BCi is the base energy cost and
Ni is the number of times instruction i is executed. The second part accounts for the circuit state
SCi_j is the energy cost when instruction i is followed by j during the program execution

Finally_

the third part accounts for energy contribution OCk of other instruction effects such as stalls and
cache misses during the program execution_ In Hsieh et al_ present a new approach_ called profile
driven program

synthesis_ to perform RT_level power estimation for high performance CPUs_ Instead of using
a macro_modeling equation to model the energy dissipation of a microprocessor_ the authors use
a synthesized program to exercise the microprocessor in such a way that the resulting instruction
trace behaves _in terms of performance and power dissipation_ much the same as the original
trace_ The new instruction trace is however much shorter than the original one_ and can hence be
simulated on a RT_level description of the target microprocessor to provide the power dissipation
results quickly_ Specifically_ this approach consists of the following steps_

Perform architectural simulation of the target microprocessor under the instruction trace
of typical application programs_

Extract a characteristic pro_le_ including parameters such as the instruction mix_ Instruction
data cache miss rates_ branch prediction miss rate_ pipeline stalls_ etc_ for the
microprocessor.

Use mixed integer linear programming and heuristic rules to gradually transform a gener
ic program template into a fully functional program_

Perform RT_level simulation of the target microprocessor under the instruction trace of the
new synthesized program _

Notice that the performance of the architectural simulator in gate vectors second is roughly to
orders of magnitude higher than that of a RT_level simulator. This approach has been applied to
the Intel Pentium processor _which is a super_ scalar pipelined CPU with _KB _way
set_associative data_ instruction and data caches_ branch prediction and dual instruction pipeline_
demonstrating _ to _ orders of magnitude reduction in the RT_level simulation time with negligible
estimation error. Behavioral_Level Power Estimation
Conversely from some of the RT_level methods that will be discussed in Section estimation
techniques at the behavioral_level cannot rely on information about the gatelevel structure of
the design components_ and hence_ must resort to abstract notions of physical capacitance and
switching activity to predict power dissipation in the design_
Information_Theoretic Models
Information theoretic approaches

for

high_level power

estimation depend on

information theoretic measures of activity .for example_ entropy_ to obtain quick power
estimates Entropy characterizes the randomness or uncertainty of a sequence of applied vectors
and thus is intuitively related to switching activity_ that is_ if the signal switching is high_ it is likely
that the bit sequence is random_ resulting in high entropy_ Suppose the sequence contains t distinct
vectors and let pi denote the occurrence probability of any vector v in the sequence_ Obviously_
The entropy of the sequence is given

where log x denotes the base logarithm of x_ The entropy achieves its maximum value of log t when
pi log pi For an n_bit vector(t,n)his makes the computation of the exact entropy very
expensive. Assuming that the individual bits in the vector are independent_ then we can write

where qi denotes the signal probability of bit i in the vector sequence. Note that this equation is
only an upperbound on the exact entropy, since the bits may be dependent. This upperbound
expression is_ however_ the one that is used for power estimation purposes. Furthermore in it has
been shown that_ under

the

temporal independence

assumption_

the average switching

activity of a bit is upper_bounded by one half of its entropy

The power dissipation in the circuit can be approximated as

Where Ctot is the total capacitance of the logic module

including gate and interconnect

capacitances_ and Eavg is the average activity of each line in the circuit which is inturn
approximated by one half of its average entropy havg. The average line entropy is computed by
abstracting information obtained from a gate_level implementation. In it is assumed that the
word_level entropy per logic level reduces quadratically from circuit inputs to circuit outputs_
whereas in it is assumed that the bit_level entropy from one logic level to next decreases in an
exponential manner. Based on these assumptions two different computational models are obtained
In Marculescu et al_ derive a closed_form expression for the average line entropy for
the case of a linear gate distribution(i.e.,)when the number of nodes scales linearly between the
number of circuit inputs n and circuit outputs m. The expression for havg is given by

where hin and hout denote the average bit_level entropies of circuit inputs and outputs_respectively_
hin is extracted from the given input sequence_ whereas hout is calculated from a quick functional
simulation

of the circuit

under

the

given

input

sequence

or

by empirical

entropy

propagation techniques for pre_characterized library modules. In Nemani and Najm propose
the following expression for havg

where Hin and Hout denote the average sectional _word_level_ entropies of circuit inputs and
outputs_ respectively_ The sectional entropy measures Hin and Hout may be obtained by monitoring
the input and output signal values during a high_level simulation of the circuit_ In practice_
however_

they are approximated as the summation of individual bit_level entropies_ hin and hout. If the circuit
structure is given_ the total module capacitance is calculated by traversing the circuit netlist and
summing up the gate loadings_ Wire capacitances are estimated using statistical wire load models_
Otherwise_ Ctot is estimated by quick mapping for example_ mapping to

input universal gates_

or by information theoretic models that relate the gate complexity of a design to the di_erence of
its input and output entropies. One such model proposed by Cheng and Agrawal in for example
estimates

This estimate tends to be too pessimistic when n is large hence in Ferrandi et al_ present
a new total capacitance estimate based on the number N of nodes i.e.,to multiplexors in the
Ordered Binary Decision Diagrams OBDD representation of the logic circuit as follows

The coefficients of the model are obtained empirically by doing linear regression analysis on
the total capacitance values for a large number of synthesized circuits. Entropic models for
the controller circuitry are proposed by Tyagi in where three entropic lower bounds on the
average Hamming distance _bit changes_ with state set S and with T states_ are provided. The
tightest lower bound derived in this paper for a sparse _nite state machine FSM i.e., tT log T
where t is the total number of transitions with nonzero steady_state probability_ is the following

where pi,j is the steady state transition probability from si to sj H(si,sj) is the Hamming distance
between the two states_ and h(pi,j) is the entropy of the probability distribution pi,j . Notice that the
lower bound is valid regardless of the state encoding used. In using a Markov chain model for the
behavior of the states of the FSM_ the authors derive theoretical lower and upper bounds for the
average Hamming distance on the state lines which are valid irrespective of the state encoding used in
the final implementation. Experimental results obtained for the mcnc benchmark suite show that
these bounds are tighter than the bounds reported.
Design Margin:
As semiconductor technology scales to the nanometer regime, the variation of process
parameters is a critical problem in VLSI design. Thus the need for variation-aware timing
analysis for the performance yield is increasing. However, the traditional worst-case corner-based
approach gives
difficult.

As

pessimistic

results,

and

makes

meeting

given

designs

specifications

an alternative to this approach, statistical analysis is proposed as a new and

promising variation-aware analysis technique. However, statistical design flow cannot be

applied easily to existing design flow, and not enough tools for statistical design exist. To
overcome these problems, new design methodology based on traditional static timing analysis
(STA) using a relaxed corner proposed

nowadays. This paper investigates the effects of corner relaxation on overall circuit
performance metrics (yield, power, area) at the gate/transistor levels. Experimental results
indicate that if we design the circuit using relaxed corner, though the circuit yield is somewhat
reduced, we can get some advantages in area and power aspects.
Reliability:
Yield and reliability are two of the cornerstones of a successful IC manufacturing technology
along with product performance and cost. Many factors contribute to the achievement of high
yield and
product

reliability,

and many

performance and

of

these also

interact

with

cost. A fundamental understanding of

failure

mechanisms and yield limitations enables the up-front achievement of these technology goals
through circuit and layout design, device design, materials choices, process optimization, and
thermo-mechanical considerations. Failure isolation and analysis, defect

analysis,

low

yield

analysis, and materials analysis are critical methodologies for the improvement of yield and
reliability. Coordination of people in many disciplines is needed in order to achieve high yield and
reliability. Each needs to understand the impact of their choices and methods on the final product.
Unfortunately, very little formal university training exists in these critical areas of IC reliability,
yield, and failure analysis.
Reliability Fundamentals and Scaling Principles

The Reliability Bathtub Curve, Its Origin and Implications

Key Reliability Functions and Their Use in Reliability Analysis

Defect Screening Techniques and Their Effectiveness

Accelerated Testing and Estimation of Useful Operating Life

Reliability Data Collection and Analysis in Integrated Circuits

Past Technology Scaling Trends

Forward Looking Projections with a Focus on Examining and Understanding of the Impact on
VLSI Reliability

Power Density Trends: Operating temperature, activation energies for dominant vlsi
failure mechanisms, and reliability impact

Reliability Strategies In Fabless Environments

Reliability of the Interconnect System

Physics and Statistics of Failure Mechanisms Associated with Interconnect

Systems

Electro-migration of Al and Cu Interconnects

Mechanical Stress Driven Metal Voiding and Cracking

Low k Materials as Interlayer Dielectrics and Their Impact on Electro-migration

Thermo-mechanical Integrity of the Interconnect System

Key Technology Parameters: Materials choices, structural and geometric effects

Extreme Scaling Impact on Wear-out Time

Technology Solutions: Alloys, metal barriers, and engineering of interfaces

Improved Electro-migration Performance under Non-DC Currents and Short Lines

Interconnect Reliability Strategies in Fabless Environments

Transistor Reliability: Dielectric Breakdown, Hot Carriers and Parametric Stability

Physics, Statistics, and Scaling Impact on Failure Mechanisms

Reliability Performance of Thin Conventional Oxides: Defects, wear-out failures

Hot Carrier Performance and Parametric Stability of P- and N-channel Devices under DC and
AC

High k Gate Dielectrics and Novel Transistor Configurations

Key Failure Mechanisms for Bipolar Transistors

Transistor Reliability Strategies in Fabless Environments

CMOS Latch-up and ESD

Physics, Scaling Impact, and Technology Dependence of CMOS Latch-up and Electrostatic
Damage (ESD)

Technology and Design Based Solutions, Device Performance, and Manufacturability

Constraints

Latch-up and ESD Assessment in Fabless Environments

Soft Errors, and Other Failure Mechanisms

Physics, Scaling Impact, and Technology Dependence of Alpha Particle and Cosmic Ray
Induced Soft Errors

Technology Solutions, Performance, and Manufacturability

Scaling:
In scaling there are really two issues
Devices

 Can we build smaller devices

What will their performance be
Wires
Try to avoid the wet noodle effect
There is concern about our ability to scale both of these Components
Limitations
Limitations to device scaling has been around since working in 3m nMOS, 22 years ago
(actually bipolar)
Worries were
Short channel effect
Punchthrough
drain control of current rather than gate
Hot electrons
Parasitic resistances
Now worries are a little different
Oxide tunnel currents
Punchthrough
Parameter control
Parasitic resistances
Transistor scaling:
People are building very short channel devices
Shown are I-V curves for 15nm L pMOS
And a short channel nMOS
The structure is strange
FinFET
But you can make them work

Wire scaling:
More uncertainty than transistor scaling
Many options with complex trade-offs
For each metal layer
Need to set H, TT, TB, e1, e2, conductivity of the metal

SPICE Tutorial:
1.Introduction
Given below is a brief introduction to simulation using HSPICE and AWAVES/Cosmoscope
in the UTD network. HSPICE is a device level circuit simulator from Synopsys. HSPICE takes a
SPICE file as input and produces output describing the requested simulation of the circuit. The
simulation output can be viewed with AWAVES (or) Cosmoscope from Synopsys. A short example
is provided to illustrate the basic procedures involved in running HSPICE.
2. Setting up your account to access HSPICE
This section shows how to setup your environment for running HSPICE.
For users who have a working CAD setup, you may just want to check that the
LM_LICENSE_FILE has

the following values in the list of all the other licenses,

/home/cad/flexlm/ti-license:/home/cad/flexlm/hspice.flx. If not, follow the procedures below:

Instructions for both bash and tcsh/csh users is provided here:
bash users:
Add the following line to the .bash_profile
LM_LICENSE_FILE=$LM_LICENSE_FILE:/home/cad/flexlm/hspice.flx ; export
LM_LICENSE_FILE

tcsh/csh users:
Add the following line to your .tcshrc
setenv LM_LICENSE_FILE ${LM_LICENSE_FILE}:/home/cad/flexlm/hspice.flx
To test if the above procedure has setup your environment successfully, invoke a new shell
(this will ensure that the new environment variables are in place). Also you will need a HSPICE input
file to test this (You can copy paste the HSPICE example given below to test this ). The input Spice
file is typically named with extension *.sp.
% hspice <your_input_file>.sp
The following message indicates trouble with invocation:
If the error is "hspice: command not found" make sure that the HSPICE
directory " /home/cad/synopsys/hspice/U-2003.09-SP1/sun58/" is included
in the $PATH variable.
Cannot execute /home/cad/synopsys/hspice/U-2003.09-SP1/sun58/hspice
or
lic: Using FLEXlm license file:
lic: /home/cad/flexlm/hspice.flx
lic: Unable to checkout hsptest
The above error may indicate that the license server maybe down, or the machine is not able
to run HSPICE.
On the other hand if the procedure was successful, you will simply see a message indicating
successful completion of simulation or errors in simulation, both of which indicate HSPICE has
run your file.
3. Setting up the HSPICE input file
Consider a self loaded min geometry inverter circuit. The objective of the HSPICE input file
below is to measure the tpLH and tpHL both graphically and otherwise. The following HSPICE file is
stored in "inv.sp". The HSPICE input file is commented adequately about the different options used
in it.
It will be beneficial to keep in mind the following differences between SPICE3 and HSPICE.

Property

SPICE3

HSPICE

Transistor

Default

If units are not specified and no SCALE statement is present, the scale

dimensions

Scale is 1u. defaults to meters. Hence for HSPICE always specify units. Eg. l=20u
Hence
depending
on

model,

with

or

without
"u".Eg. l=20
Input
Pattern

bit

PBIT

or Only PWL format is supported. Howevere to convert a PBIT(Bit stream

PWL

format) to a PWL form, you can use the script and help at the following
page:
http://www.utdallas.edu/~poras/courses/ee6325/lab/hspice/pbit2pwl.ht ml

Output

Print

format

Punch

and SIMG does not read HSPICE output, only AWAVES or Cosmoscope
files can read HSPICE output. Also a graphical output produced only if
.option post=1 is provided. The .print command is of no consequence to

produced

if graphical output.

only

requested as
.punch/.print.
SIMG

reads

the .pun file

Line

In order to In order to specifying a continuing line '+' character is used at the start of

continuation specifying a the second line.

continuing

'&' + 0ns pvdd 1ns pvdd

line

character is
used at the
end of the
first line.
Eg: Vin in
gnd PWL&
0ns

Eg: Vin in gnd PWL

pvdd

1ns pvdd

HSPICE Example File:

* Self loaded min geometry inverter, sample HSPICE file
* Include the model files
* Include the hspice model files for 0.18u technology.
.include
/home/cad/vlsi/models/hspice/cmos0.18um.model
*********************************************************************
* The subcircuit for the inverter
.subckt invert in out vdd gnd
.param length=0.2u
m01 out in vdd vdd pfet w='4*length' l='length'
m02 out in gnd gnd nfet w='1.5*length' l='length'
.ends
*********************************************************************
* The main inverter
X1 in out vdd gnd invert
* Four loads for the inverter
X2 out out1 vdd gnd invert
X3 out out2 vdd gnd invert
X4 out out3 vdd gnd invert
X5 out out4 vdd gnd invert
* PWL pattern for the input, represents a bit stream 1100101
* Slew=1ns, bit time=5ns
Vin in gnd PWL 0ns pvdd 1ns pvdd 5ns pvdd 6ns pvdd 7ns 0 10ns 0
+ 15ns 0 16ns pvdd 21ns pvdd 22ns 0 25ns 0 26ns pvdd
* Parametric definitions

.param pvdd=2.0v
* Power supplies
vvdd vdd 0 pvdd
vgnd gnd 0 0
* Control statements
.option post=1
.TR 0.05ns 30ns
.print TR V(in out)
* Measure statements help in calculating TPLH, TPHL etc, without
* opening the waveform viewer
.measure tran tplh trig v(in) val='0.5*pvdd' fall=1 targ v(out) val='0.5*pvdd' rise=1
.measure tran tphl trig v(in) val='0.5*pvdd' rise=1 targ v(out) val='0.5*pvdd' fall=1
.END

4. Running HSPICE simulations

The following commands can be used to simulate the above HSPICE file stored in inv.sp and store all
the simulation results with file prefix as "inv"
% hspice inv.sp -o inv
This results in the creation of the following output files:
inv.ic -> Operating point node voltages (initial conditions)
inv.lis -> Output listing
inv.mt0 -> Transient analysis measurement results
inv.pa0 -> Subcircuit cross-listing
inv.st0 -> Output status
inv.tr0 -> Transient analysis results
5. Analyzing the outputs

In the above example, the output data can be analyzed both graphically as well as in text form.
Text outputs:
To view the results of the .measure computation, execute:
% cat inv.mt0
$DATA1 SOURCE='HSPICE' VERSION='2003.09-SP1'
.TITLE ' '
tplh

tphl

temper

alter#

3.416e-10 1.002e-09 25.0000 1.0000

As can be seen above, the values of propogation delay have been obtained even before the
waveform analysis software has been opened.
Graphical outputs:
I. Synopsys Awaves:
To invoke AWAVES run the following command:
% awaves
If you get the error "awaves: command not found" make sure that the AWAVES directory
"/home/cad/synopsys/hspice/U-2003.09-SP1/sun58/" is included in the $PATH variable.
Once

invoked,

open

the

design

using

the

pull

down

menu

options:

Design->Open and select inv.sp and then highlight the tr0 (Transient response) item in the select box.
You will also see the hierarchy of the netlist and the types of analysis and the individual signals
in separate lists in the window. Select Hierarchy -> Top, Types -> Voltages and select the voltages
you want to observe For eg. in and out by double clicking on the names. You will see the screen
below for the stimulus provided in inv.sp.

II. Synopsys Cosmoscope:

To invoke Cosmoscope run the following command:
%cscope
If you get the error "awaves: command not found" make sure that the AWAVES directory "
/home/cad/synopsys/cosmo/ai_bin/" is included in the $PATH variable.
Once

invoked,

open

the

design

using

the

pull

down

menu

options:

File -> Open -> Plotfiles and select file inv.tr0 in the working directory. A Signal Manager window
and signal window opens. Select the necessary signals to be plotted by double-clicking them.For
example v(in) and v(out) by double clicking on the signal names in the signal window. You will see
the

screen

below

for

the

stimulus

provide

d in inv.sp.
Device models:
The motivation for this investigation stems from three main concerns:
1. The usual parameterization of device models for device and circuit simulation causes
problems due to the interdependence of the parameters. It is not physically realistic to change any
one parameter without determining the change in the process technology that would produce
such a change in the parameter. Then all the other parameters which also depend on this
change in the technology must be adjusted accordingly. In addition, it is quite difficult to determine
the effect of a specific change in a new technology since the available parameters each depend
on a number of technology parameters.
2. The predictive performance of present models is not good. It has usually been necessary
to fabricate devices in any chosen technology, and extract parameters, and then fit the model to
this specific technology by use of additional adjustment parameters. Of course, this
procedure is

reasonable and useful once a technology has been chosen. However, it would be useful if the model
could produce fairly accurate results if only the process specifications are used. Without such
predictive accuracy it is difficult to make an initial choice of technology.
3. Most models have been developed for digital applications where devices operate above
threshold and therefore are not strongly temperature sensitive. This causes problems for
modeling analog circuits which use subthreshold operation. In particular, the temperature
dependence of subthreshold behavior has not been fully explored. In many models some
parameters which are temperature dependent have been assumed to be constant. Device and circuit
models are all based on the physical properties of semiconductor materials, the dimensions of the
devices, and on theoretical and empirical equations which are intended to model electrical
behavior. The distinction between theoretical and empirical equations is often unclear. Most of the
equations are substantially empirical. Of all the equations, one of the most fundamental, and
problematic, is the equation for ni, the intrinsic carrier concentration of a semiconductor. The
definition of ni derives from the

thermodynamic

equilibrium of electron and hole formation, based on the fact that the energy gap is a Gibbs energy.
The equilibrium equation is

,
where n is the electron concentration, p is the hole concentration, Nc is the density of effective states
in the conduction band, Nv is the density of effective states in the valence band, Eg is the band gap, k
is Boltzmanns constant and T is the absolute temperature. The carrier concentration is then given by
It would appear to be a simple matter to substitute Si values for Nc, Nv, Eg, and the value of the
constant kT to obtain an accurate value of ni. However, the theoretical and experimental knowledge
required for accurate values of Nc and Nv is even now incomplete. In the early 1960s, when Sibased circuits were beginning to be designed and fabricated very little was known about Nc and
Nv, but estimates were required for practical use. This led to approximations based on work.
The key approximation was that chosen by Grove in . This approximation is the still widely used

Determination of Intrinsic Carrier Concentration (ni)

The empirical expressions for Eg from Bludau and for Nc and Nv from Sproul and Green
provide the values needed for equation (2.1). Our derivation of the equation for ni follows Green.
Therefore, the exciton binding energy term, using the value Exb = 14.7meV, is included in the
band gap, Eg. In the past, this term has either been neglected, or in some cases a value of 10meV
has been

used. Green [9] and Sproul and Green are by far the best references for the history, theory, and
experimental measurements leading to reliable values for ni. Thus, the equations are:

Circuit and device characterization:

The modeling procedure is introduced in this chapter, taking into consideration the
requirements for a good MOSFET analog model, discussed in the previous chapter. We note here two
main aspects of our modeling approach;
a. The model must describe accurately all the operating regions in order to be
integrated in a circuit simulator.
b. The current, conductance, and transconductance must be continuous in all regions
of operation.
Our main goal in this chapter is to determine the drain current for any combination of
terminal voltages. The chapter is divided into two main parts. Throughout the first part, it is assumed
that the channel is sufficiently long and wide, so that edge effects are confined to a negligible part of
it. While in the second part we incorporate the short and narrow channel effects to the model. We also
assume that the substrate is uniformly doped. The doping concentration will be assumed to be p-type
and the modification to non uniform doping will be discussed later in this chapter.
Gradual Channel approximation (GCA)

Figure 2.1 The MOSFET structure.

Analytical or semi-analytical modeling of MOSFET characteristics is usually based on the so-called

Gradual Channel Approximation (GCA). In this approximation, we assume that the gradient of the
electric field in the y direction, F/y is much smaller than the gradient of the electric field in the x
direction ,F/x. Which enable us to determine the inversion and depletion
charge densities under the gate in terms of a one-dimensional electrostatic problem for the
direction perpendicular to the channel. By applying of the two dimensional Poisson's
equation for the semiconductor, refer to Fig. 2.1 region (2),

(2.2.1)
if we assume that the GCA is valid equation 2.2.1 may be approximated to the following one
dimensional differential equation

(2.2.2)
we approach the source and drain junctions, the GCA becomes invalid (Fig. 2.1 regions (1) and (3))
because of the increasing longitudinal field due to the pn junctions which make F/y comparable or
even larger than F/x . However, for the long channel MOSFET's these transition regions can be
neglected with respect to the total length of the device. In order to account for the effect of these
regions, it is necessary to use two dimensional analysis requiring a numerical solution of 2.2.1.
Validity of the GCA
The validity of GCA can be checked by making rough estimates of the variation in the
longitudinal and vertical field components. We will establish expressions that allow the GCA to be
checked under strong inversion.

(2.2.3)
For a MOSFET at 300K with L = 1.0m, tox = 30 nm, VGS-VT = 0.5, and VDS = 0.5 V, the left
hand side of inequality 2.2.3 is 2300, indicating that the the GCA is a very good approximation
for
such a MOSFET. This also implies that the GCA can be valid even in submicron MOSFETs,
provided
that VGSVT is not too small.
2.3 The long channel current model
The derivation of the dc drain current relationship recognizes that, in general, the current
in the channel of a MOSFET can be caused by both drift and diffusion current. In an NMOSFET
we may assume the following resonable approximation :
i- The drain current is mainly carried by electrons .
ii- The current flows almost in the y direction.

iii- No sources or sinks in the channel.

Note that in weak inversion the surface potential along the channel in long channel MOSFETs is
almost constant. Thus Fy/y is very small, implying that Fy/y<<Fx/x. T hus in long
channel MOSFET the GCA is valid both in strong and weak inversion regions . Which enable us to
reach to the following general relationship of drain current.
General Drift-Diffusion current equation in MOSFET:
This is the drift-diffusion drain current of the form

(2.3.1)
where n is the electron surface mobility in the channel, W is the channel width, Qi is the inversion
charge density per unit area, c is the quasi Fermi potential (the difference between Efn at the
surface of the semiconductor a d Efp in the bulk of the semiconductor), s is the surface potential
referenced to the bulk potential, and t is the thermal voltage (=kT/q). The first term is the
drift

current component, while the second term is the diffusion current component. In both

components, n is the electrons' surface mobility being less than the mobility in the bulk due to
surface scattering.
Voltage-Charge equation from the Transverse electric field:
In order to eliminate the electron charge density Qi term in the current charge equation, a
second relationship is required that relates the electron charge density to the applied potentials. Using
the relationship between voltage and charge appearing across the MOS capacitor we have
Cox (VG - ms - s )= -(Qi+QB+Qox+Qit ) (2.3.2)
where VG is the gate voltage referenced to the bulk potential, ms is the metal semiconductor
work function difference, QB is the depletion (bulk impurity) charge density per unit area, Qox is
the sum of the effective net oxide charge per unit area at the Si-SiO2 interface, and Qit is the
interface trapped charge density per unit area. Different approximations have been introduced in
order to express the different MOS charges (QB, Qox, Qit) in terms of the applied voltages, then
using eq. (2.3.2) to compute the inversion charge density Qi. The resulting charge is then used in eq.
(2.3.1) to determine the drain current; Four main approaches then follow, after them we shall discuss
the proposed approach recently developed in ICL1 and modified by this work.
Interconnect simulation models:
The classical long-channel Pao and Sah model
The Pao-Sah model [11,14], published in 1966, was the first advanced long channel MOSFET
model to be developed. While it retained the GCA, it didnt invoke the depletion approximation and
permitted carrier transport in the channel by both drift and diffusion current. The formulation of
the

drain current equation is therfore general, but as a result requires numerical integration in
two dimensions, which limits its application in CAD tools.
Approximations:
i . Gradual Channel Approximation is used.
ii . Constant mobility is assumed.
iii . Uniform substrate doping is considered.
Advantages:
i . It is physically based.
ii . It gives a continuous representation of the device characteristics from weak to
strong inversion even to the saturation mode of operation.
Disadvantages:
i. It requires excessive computational requriments since it requires numerical integration
in two dimension, rendering it unsuitable to be used for circuit CAD.

The charge-sheet based models

The limited practical utility of the Pao-Sah model motivated a search for an approximate
advanced analytical model, that is still accurate over a wide range of operating conditions. The
charge sheet model, introduced separately by Bacarani and Brews in 1978, has become the most
widely adopted long channel MOSFET model that is accurate over the entire range of inversion.
In this model the inversion layer is supposed to be a charge sheet of infinitesimal thickness
[11,15,16] (charge sheet approximation). The inversion charge density Qi can then be calculated in
terms of the surface potential s. The drain current (2.3.1) is then expressed in terms of the surface
potential at the source and drain boundaries of the channel.
Approximations:
i . Gradual Channel Approximation is used.
ii . The mobility is assumed to be proportional to the electric field and is constant with
position along the channel.
iii . Uniform substrate doping is considered.
Advantages:
i . It is physically based.
ii . It gives a continuous representation of the device characteristics from weak to strong
inversion even to the saturation mode of operation.
ii . The charge sheet approximation introduces negligibly small error, and it is
more computationally efficient than the classical model.
Disadvantages:

i . The boundary surface potentials cannot be expressed explicitly in terms of the bias
voltages applied to the device, but must be found by a numerical process.
ii . The model is not valid in depletion or accumulation.
Different approaches have been introduced to circumvent this disadvantage. In it is shown that
accurate numerical solutions for these surface potentials can be obtained with negligible
computation time penalty. In the surface potentials are computed using cubic splines functions.
In the implicit equation including the surface potential is replaced by an approximate function.
Although all of these approaches have given good results, they have neglected the effect of the
interface trap charge which is important in determining the subthreshold characteristics of the
device, namely the subthreshold swing (the gate voltage swing needed to reduce the current by one
decade).
Bulk Charge Model
The Bulk Charge model also known as variable depletion charge model, was developed in
1964, describes the MOSFET drain current only in strong inversion but of course has less
computational requirements.
Approximations :
i . Drift current component only is considered
ii . Constant surface potential is assumed
iii . Id considered zero below threshold
Advantages :
i . Less computational time than the charge sheet model
Disadvantages :
i . The subthreshold region not defined

Square law model

This model has great popularity, when a first estimate to device operation, or simulating
a circuit with a large number of devices is required. This model is obtained from the bulk charge
model, on the assumption that VDS << 2f+VBS .
i . Drift current component only is considered
ii . Constant surface potential is assumed
iii . Id considered zero below threshold
iv . VDS << 2f+VBS
Advantages :
i . Very small computational time than any other model
ii . Suitable for hand calculations
Disadvantages :
i . The subthreshold region is not defined

ii . Overestimates the drain current in saturation region

Approximate models
There exists a large number of introduced approximate models. All of these models
originate from Brews' charge sheet model, where approximations to the surface potentials in various
operating regions of the device have been used. This leads to different current equations each
valid only in a specific region. The resulting equations are then empirically joined using
different mathematical conditions of continuity.
Advantages:
i . They have good accuracy in the desired region of operation.
ii . They are very efficient from the point of view of computational time.
Disadvantages:
i . The error increases in the transition regions between different modes of
operations. ii . They include many non-physical fitting parameters.
Modified charge sheet model
The last discussed MOSFET models, have a common illness, no interface charges are
included which play a great role in subthreshold region. So a modified model to the charge sheet
model, which include the effect of interface charges is carried out in ICL, and will be presented now.
The derivation begins by rewriting equation (2.3.1) in the following form :
I D = I D1+ I D2 (2.3.6.1)
where ID1 is due to the presence of drift:

2.3.6.2)
and ID2 is due to the presence of diffusion:

(2.3.6.3)
after mathematical manipulation and following the same approximations as charge sheet model
we reach the following drain current equations:

(2.3.6.4)

(2.3.6.5)
where s0 is the surface potential at the source end of the channel, sL is the surface potential at the
drain end of the channel, both are referred to the bulk. And their values are computed from the
following two implicit equations.

(2.3.6.6)

(2.3.6.7)

UNIT III COMBINATIONAL AND SEQUENTIAL CIRCUIT DESIGN

Circuit families and its comparison:
The method of logical effort does not apply to arbitrary transistor networks, but only to
logic gates. A logic gate has one or more inputs and one output, subject to the following restrictions:
_ The gate of each transistor is connected to an input, a power supply, or the output; and _ Inputs are
connected only to transistor gates. The first condition rules out multiple logic gates masquerading
as one, and the second keeps inputs from being connected to transistor sources or drains, as
in transmission gates without explicit drivers.

Pseudo-NMOS circuits
Static CMOS gates are slowed because an input must drive both NMOS and PMOS
transistors. In any transition, either the pullup or pulldown network is activated, meaning the input
capacitance of the inactive network loads the input. Moreover, PMOS transistors have poor
mobility and must be sized larger to achieve comparable rising and falling delays, further
increasing input capacitance. Pseudo-NMOS and dynamic gates offer improved speed by
removing the PMOS transistors from loading the input. This section analyzes pseudo-NMOS
gates, while section 10.2 explores dynamic logic. Pseudo-NMOS gates resemble static gates,
but replace the slow PMOS pullup stack with a single grounded PMOS transistor which acts as a
pullup resistor. The effective pullup resistance should be large enough that the NMOS transistors
can pull the output to near ground, yet low enough to rapidly pull the output high. Figure 10.1
shows several pseudo-NMOS gates ratioed such that the pulldown transistors are about four
times as strong as the pullup. The analysis presented in Section 9.1 applies to pseudo-NMOS
designs. The logical effort follows from considering the output current and input capacitance
compared to the reference inverter from Figure Sized as shown, the PMOS transistors produce 1/3
of the current of the reference inverter and the NMOS transistor stacks produce 4/3 of the current
of the reference inverter. For falling transitions, the output current is the pulldown current minus the
pullup current which is fighting the pulldown,For rising transitions, the output current

PSEUDO-NMOS CIRCUITS

is just the pullup current, 1/3. The inverter and NOR gate have an input capacitance of 4/3. The
falling logical effort is the input capacitance divided by that of an inverter with the same output
current, or The rising logical effort is three times greater, because the current produced on a rising
transition is only one third that of a falling transition. The average logical effort is g = (4=9+4=3)=2
= 8. This is independent of the number of inputs, explaining why pseudo-NMOS is a
way to build fast wide NOR gates. Table 10.1 shows the rising, falling, and average logical efforts of
other pseudo-NMOS gates, assuming _ = 2 and a 4:1 pulldown to pullup strength ratio. Comparing
this with Table 4.1 shows that pseudo-NMOS multiplexers are slightly better than CMOS
multiplexers and that pseudo-NMOS NAND gates are worse than CMOS NAND gates. Since
pseudo- NMOS logic consumes power even when not switching, it is best used for critical NOR
functions where it shows greatest advantage. Similar analysis can be used to compute the logical
effort of other logic technologies, such as classic NMOS and bipolar and GaAs. The logical
efforts should be normalized so that an inverter in the particular technology has an average logical
effort of 1.
10.1.1 Symmetric NOR gates
Johnson [4] proposed a novel structure for a 2-input NOR, shown in Figure 10.2. The gate
consists of two inverters with shorted outputs, ratioed such that an inverter pulling down can
overpower an inverter pulling up. This ratio is exactly the same as is used for pseudo-NMOS gates.
The difference is that when the output should rise, both inverters pull up in parallel, providing
more current than is available from a regular pseudo-NMOS pullup. The input capacitance of each
input is
2. The worst-case pulldown current is

equal to that of a unit inverter, as we had found in the analysis of pseudo-NMOS NOR gates.
The pullup current comes from two PMOS transistors in parallel and is thus 2=3 that of a unit
inverter. Therefore, the logical effort is 2=3 for a falling output and 1 for a rising output. The
average effort is g = 5=6, which is better than that of a pseudo-NMOS NOR and far superior to that
of a static CMOS NOR! and even for NAND gates. Exercises 10-3 and 10-4 examine the design
and logical effort of such structures.
Domino circuits
Pseudo-NMOS gates eliminate the bulky PMOS transistors loading the inputs, but pay the
price of quiescent power dissipation and contention between the pullup and pulldown transistors.
Dynamic gates offer even better logical effort and lower power consumption by using a clocked
precharge transistor instead of a pullup that is always conducting. The dynamic gate is precharged
HIGH then may evaluate LOW through an NMOS stack. Unfortunately, if one dynamic inverter
directly drives another, a race can corrupt the result. When the clock rises, both outputs have been
precharged HIGH.

The HIGH input to the first gate causes its output to fall, but the second gates output also
falls in response to its initial HIGH input. The circuit therefore produces an incorrect result
because the second output will never rise during evaluation, as shown in Figure 10.3. Domino
circuits solve this problem by using inverting static gates between dynamic gates so that the input to
each dynamic gate is initially LOW. The falling dynamic output and rising static output ripple
through a chain of gates like a chain of toppling dominos. In summary, domino logic runs 1:5 to 2
times faster than static CMOS logic [2] because dynamic gates present a much lower input
capacitance for the same output current and have a lower switching threshold, and because the
inverting static gate can be skewed to favor the critical monotonically rising evaluation edges.
Figure shows some domino gates. Each domino gate consists of a dynamic gate followed by an
inverting static gate1. The static gate is often but not always an inverter. Since the dynamic gates
output falls monotonically during evaluation, the static gate should be skewed high to favor its
monotonically rising output. We have calculated the logical effort of high-skew gates in Table 9.2
and will compute the logical effort of dynamic gates in

the next section. The logical effort of a domino gate is then the product of the logical effort of the
dynamic gate and of the high-skew gate. Remember that a domino gate counts as two stages when
choosing the best number of stages. A dynamic gate may be designed wit h or without a
clocked evaluation transistor; the extra transistor slows the gate but eliminates any path between
power and ground during precharge when the inputs are still high. Some dynamic gates include
weak PMOS transistors called keepers so that the dynamic output will remain driven if the
clock stops high. Domino designers face a number of questions when selecting a circuit topology.
How many stages should be used? Should the static gates be inverters, or should they perform
logic? How should precharge transistors and keepers be sized? What is the benefit of removing
the clocked evaluation transistors? We will show that domino logic should be designed with a stage
effort of 22:75, rather than 4 that we found for static logic. Therefore, paths tend to use more
stages and it is rarely beneficial to perform logic with the inverting static gates.
Logic power logic design:
Is to Reduce dynamic power and static power in a circuit

a:

C:

VDD:

f:

Reduce static power,Reduce dynamic power

a: clock gating, sleep mode

C: small transistors (esp. on clock), short wires

VDD:

f:

Reduce static power, Reduce dynamic power

a: clock gating, sleep mode

C: small transistors (esp. on clock), short wires

VDD: lowest suitable voltage

f: lowest suitable frequency

Reduce static power

Selectively use ratioed circuits

Selectively use low Vt devices

Leakage reduction: stacked devices, body bias, low temperature.

Circuit design of latches and flip flops:

Another class of logic circuits are sequential circuits. These circuits are two-valued networks
in

which the outputs at any instant are dependent not only upon the inputs present at that instant but also
upon the past history (sequence) of inputs.
Sequential circuits are classified into:
The block diagram of a sequential circuit is shown below:

Synchronous sequential circuits Their behaviour is determined by the values of the

signals at only discrete instants of time.

Asynchronous sequential circuits Their behaviour is immediately affected by the

input signal changes.

The basic logic element that provides memory in many sequential circuits is the flipflop.

1. Latches
Latches form one class of flip-flops. This class is characterized by the fact that the timing of the
output changes is not controlled. Although latches are useful for storing binary information and
for the design of asynchronous sequential circuits, they are not practical for use in synchronous
sequential circuits.
1.2 The SR Latch
It is a circuit with two cross-coupled NOR gates or two cross-coupled NAND gates. The one
with
NOR gates is shown below:

The condition that is undefined is when both inputs are equal to 0 at the same time.
Comparing the NAND with the NOR latch note that the input signals for the NAND require the
complement of those values used for the NOR latch. Because the NAND latch requires a 0 signal
to change its state, it is sometimes referred to as an S-R latch. The operation of the SR latch can
be modified to provide an additional control input that determines when t he state of the latch can
be

changed. An SR latch with control input is shown below: The control input C acts as an enable signal
form the other two inputs. An indeterminate condition occurs when all three inputs are equal to 1.
This condition makes the circuit difficult to manage and is seldom used in practice. Nevertheless, it
is an important circuit because other latches and flip-flops are realized from it.

The Gated D Latch

One way to eliminate the undesirable condition of the indeterminate state in t he SR latch is
to ensure that inputs S and R are never equal to 1 at the same time. This is done by the D latch:

Flip-Flops
When latches are used for the memory elements in sequential circuits, a serious difficulty
arises.

Recall

that

latches

have the property

of

immediate

output

responses

(i.e.,

transparency). Because of this the output of a latch cannot be applied directly (or through logic) to
the input of the same or another latch when all the latches are triggered by a common clock
source. Flip -flops are used to overcome this difficulty.

Edge-Triggered D Flip-Flop
A D flip-flop may be realized with two D latches connected in a master-slave configuration:
The circuit samples the D input and changes its Q output only at the negative-edge of the controlling
clock signal (CLK). It is also possible to design the circuit so that the flip-flop output changes on the
positive edge of the clock (transition from 0 to 1). This happens in a flipflop that has an additional
inverter between the CLK terminal and the junction between the other inverter and input C on the
master latch. An efficient realization of a positive edge-triggered D flip-flop uses three SR latches:

The graphic symbol for the edge-triggered D flipflop is:

Other Flip-Flops
The most economical and efficient flip-flop in terms of transistor count and silicon area is
the D flip-flop. Other types of flip-flops can be realized by using the D flip-flop and external logic.
Two flipflops widely used in the design of digital systems are the JK and the T flip-flops. There are
three operations that can be performed with a flip-flop: set it to 1, reset it to 0, complement its
output. The JK flip-flop performs all three:

Characteristic Tables
A characteristic table defines the logical properties of a flip-flop by describing its operation
in tabular form. The flip-flops characteristic tables are:

Q(t) refers to the present state prior to the application of a clock edge.
Q(t + 1) is the next state one clock period later. The clock edge input is not included in the
characteristic tables, but is implied to occur between time t and t + 1.
Characteristic Equations
The logical properties of a flip-flop as described in its characteristic table can be
expressed also algebraically with a characteristic equation. For the D flip-flop the characteristic
equation is:
Q(t +1) = D
It states that the next state of the output will be equal to the value of input D in the present
state.
The characteristic equation for the JK flip-flop is:

where Q is the value of the flip-flop output prior to the application of a clock edge. Finally, the
characteristic equation for the T flipflop is:

Static sequencing element methodology:

Power has become critical metric and key differentiator in sub-65nm SOC designs, due
to growing power density driven by technology scaling and chip integration. This tutorial
provides overview of the low-power design methodologies and techniques in production
SOC design perspective, emphasizing on the real design considerations and impact on chip
success. We shall discuss pros and cons of the methods and techniques considering impacts on
chip design schedule, yield, and overall power-performance target. We shall also provide
design guidance and

recommendations in various design steps and decision making points, based on our years of
successful experience in production low-power SOC designs.
This tutorial is organized in two parts. In the first part, we shall overview power related
challenges in sub-60nm SOC design and state-of-the-art techniques to reduce chip power. We shall
give a holistic view from chip level to system and application levels. Practical industrial examples
will be used to show how power savings can be achieved in modern SoC, processors and computer
systems. In the second part, we shall describe production low-power design methodology and
techniques particularly the power-gating and the voltage/frequency scaling which are the two
advanced power reduction methods used effectively in sub-65nm production low-power designs. We
shall explain when, where and how these methods and techniques are applied to a chip according to
the design goals and time-to-market requirement. We shall also cover production low-power
design methodology and flow with UPF power intent and unified design environment.
Sequencing dynamic circuits:
Simulation-based techniques for dynamic compaction of test sequences are proposed.
The first technique uses a fault simulator to remove test vectors from the partially-specified test
sequence generated by a deterministic test generator if the vectors are not needed to detect the
target fault, considering that the circuit state may be known. The second technique uses genetic
algorithms to fill the unspecified bits in the partially-specified test sequence in order to increase the
number of faults detected by the sequence. Significant reductions in test set sizes were observed
for all benchmark circuits studied. Fault coverages improved for many of the circuits, and execution
times often dropped as well, since fewer faults had to be targeted by the computationintensive deterministic test generator. 1 Introduction Deterministic test generators for single stuckat faults in sequential circuits typically target individual faults, and once a test is generated, the
test is fault simulated. CMOS inverters (Complementary NOSFET Inverters) are some of the
most widely used and adaptable MOSFET inverters used in chip design. They operate with very
little power loss and at relatively high speed. Furthermore, the CMOS inverter has good logic buffer
characteristics, in that, its noise margins in both low and high states are large.
This short description of CMOS inverters gives a basic understanding of the how a CMOS
inverter works. It will cover input/output characteristics, MOSFET states at different input
voltages, and power losses due to electrical current. A CMOS inverter contains a PMOS and a NMOS
transistor connected at the drain and gate terminals, a supply voltage VDD at the PMOS source
terminal, and a ground connected at the NMOS source terminal, were VIN is connected to the
gate terminals and VOUT is connected to the drain terminals.(See diagram). It is important to notice
that the CMOS does not contain any resistors, which makes it more power efficient that a
regular resistor-MOSFET

inverter.As the voltage at the input of the CMOS device varies between 0 and 5 volts, the state of the
NMOS and PMOS varies accordingly. If we model each transistor as a simple switch activated by
VIN, the inverters operations can be seen very easily.

UNIT IV CMOS TESTING

Need for testing:
Iddq testing is a method for testing CMOS integrated circuits for the presence of manufacturing
faults.
It relies on measuring the supply current (Idd) in the quiescent state (when the circuit is not switching
and inputs are held at static values). The current consumed in the state is commonly called Iddq for
Idd (quiescent) and hence the name. Iddq testing uses the principle that in a correctly operating
quiescent CMOS digital circuit, there is no static current path between the power supply and ground,
except for a small amount of leakage. Many common semiconduct or manufacturing faults will
cause the current to increase by orders of magnitude, which can be easily detected. This has the
advantage of checking the chip for many possible faults with one measurement. Another advantage is
that it may catch faults that are not found by conventional stuck-at fault test vectors. Iddq testing
is somewhat more complex than just measuring the supply current. If a line is shorted to Vdd, for
example, it will still draw no extra current if the gate driving the signal is attempting to set it to
'1'. However, a different vector set that attempts to set the signal to 0 will show a large increase in
quiescent current, signalling a bad part. Typical Iddq test vector sets may have 20 or so vectors.
Note that Iddq test vectors require only controllability, and not obser vability. This is because the
observability is through the shared power supply connection. Iddq testing has many advantages:

It is a simple and direct test that can identify physical defects.

The area and design time overhead are very low.

Test generation is fast.

Test application time is fast since the vector sets are small.

It catches some defects that other tests, particularly stuck-at logic tests, do not.

Drawback: Compared to scan testing, Iddq testing is time consuming, and then more expensive,
since is achieved by current measurements that take much more time than reading digital pins in
mass production.

Testers and test programs:

As processes shrink (see Moore's la w), the leakage current becomes much higher and less
predictable. This makes it difficult to tell a low leakage part with a defect from a naturally high
leakage part. Also, increasing circuit size means a single fault will have a lower percentage effect,
making it harder for the test to detect. However, Iddq is so useful that designers are taking steps to
keep it working. One particular technique that helps is power gating, where the entire power
supply to each block can be switched off using a low leakage switch. This allows each block to
be tested individually or in combination, which makes the tests much easier when compared to
testing the whole chip.
Text fixtures:
In generic xUnit, a test fixture is all the things that must be in place in order to run a test and expect a
particular outcome. Frequently fixtures are created by handling setUp() and tearDown() events of
the unit testing framework. In setUp() one would create the expected state for the test, and in
tearDown() it would clean up what had been set up.
Four phases of a test:
1. Set up -- Setting up the test fixture.
2. Exercise -- Interact with the system under test.
3. Verify -- Determine whether the expected outcome has been obtained.
4. Tear down -- Tear down the test fixture to return to the original state.
Use of fixtures
Some advantages of fixtures include separation of the test initialization (and destruction) from the
testing, reusing a known state for more than one test, and special assumption by the testing
framework that the fixture set up works.
Design for testability:
VLSI designers have a wide variety of CAD tools to choose from, each with their own
strengths and weaknesses. The leading Electronic Design Automation (EDA) companies include
Cadence, Synopsys, Magma, and Mentor Graphics. Tanner also offers commercial VLSI design
tools. The leading free tools include Electric, Magic, and LASI. This set of laboratories uses the
Cadence and Synopsys tools because they have the largest market share in industry, are capable of
handling everything from simple class projects to state-of-the-art integrated circuits. The full set
of tools is extremely expensive but the companies offer academic programs to make the tools
available to

universities at a much lower cost. The tools run on Linux and other flavors of Unix. Setting up and
maintaining the tools involves a substantial effort. Once they are setup correctly, the basic tools are
easy to use, as this tutorial demonstrates. Some companies use the Tanner tools because their list
price is much lower and they are easy to use. However, their academic pricing is comparable with
Cadence and Synopsys, giving little incentive for universities to adopt Tanner. The Electric
VLSI Design System is an open-source chip design program developed by Electric presently
does not read the design rules for state-of-the-art nanometer processes and poorly integrates with
synthesis and place & route. Magic is a free Linux-based layout editor with a powerful but awkward
interface that was once widely used in universities. The Layout System for Individuals, LASI,
developed by David Boyce, is freely available and runs on Windows. It was last updated in 1999.
There are two general strategies for chip design. Custom design involves specifying how every
transistor is connected and physically arranged on the chip.

Synthesized design involves

describing the function of a digital chip in a hardware description language such as Verilog or
VHDL, then using a computer-aided design tool to automatically generate a set of gates that perform
this function, place the gates on the chip, and route the wires to connect the connect the gates. The
majority of commercial designs are synthesized today because synthesis takes less engineering time.
However, custom design gives more insight into how chips are built and into what to do when
things go wrong. Custom design also offers higher performance, lower power, and smaller chip
size. The first two labs emphasize the fundamentals of custom design, while the next two use logic
synthesis and automatic placement to save time.
II. Tool Setup
These labs assume that you have the Cadence and Synopsys tools
installed.
The tools generate a bunch of random files. Its best to keep them in one place. In
your
home directory, create some directories by typing:
mkdir IC_CAD
mkdir IC_CAD/cadence
III. Getting Started
Before you start the Cadence tools, change into the cadence
directory:
cd ~/IC_CAD/cadence
Each of our tools has a startup script that sets the appropriate paths to the tools and invokes them.
Start Cadence with the NCSU extensions by running cad-ncsu & A window labeled icfb will open up.
This is the Integrated Circuit Front and Back End (e.g. schematic and layout) software, part
of Cadences Design Framework interface. A Whats New and a Library Manager window may
open up too. Scroll through the icfb window and look at the messages displayed as the tool loads up.
Get in the habit of watching for the messages and recognizing any that are out of the ordinary. This
is very helpful when you encounter problems.

All of your designs are stored in a library. If the Library Browser doesnt open, choose Tools
Library Manager. Youll use the Library Manager to manipulate your libraries. Dont try to move
libraries around or rename them directly in Linux; there is some funny behavior and you are likely
to break them. Familiarize yourself with the Library Manager. Your cds.lib file includes many
libraries from the NCUS CDK supporting the different MOSIS processes. It also includes libraries
from the University of Utah. The File menu allows you to create new libraries and cells within a
library, while the Edit menu allows you to copy, rename, delete, and change the access permissions.
Create a library by invoking File New Library in the Library Manager. Name the library
lab1_xx, where xx are your initials. Leave the path blank and it will be put in your current working
directory (~/IC_CAD/cadence). Choose the Attach to existing tech library and accept the default,
UofU AMI 0.60u C5N (3M, 2P, high-res). This is a technology file for the American
Microsystesm (now Orbit Semiconductor) 0.6 m process, containing design rules for layout.
IV. Schematic Entry
Our first step is to create a schematic for a 2-input NAND gate. Each gate or
larger component is called a cell. Cells have multiple views. The schematic view for a cell built with
CMOS transistors will be called cmos_sch. Later, you will build a view called layout specifying how
the cell is physically manufactured. In the Library Manager, choose File New Cell View In your
lab1_xx library, enter a cell name of nand2 and a view name of cmos_sch. The tool should be
Composer - Schematic. You may get a window asking you to confirm that cmos_sch should be
associated with this tool. The schematic editor window will open. Your goal is to draw a gate like
the one shown in Figure 1. We are working in a 0.6 m process with = 0.3 m. Unfortunately, the
University of Utah technology file is configured on a half-lambda grid, so grid units are 0.15
m. Take care that everything you do is an integer multiple of so you dont come to grief later on.
Our NAND gate will use 12 (3.6 m) nMOS and pMOS transistors.
Choose Add Instance to open a Component Browser window. (The menu lists the keyboard
shortcut for each command, such as i for add instance. Youll want to learn the shortcuts you use
most often.) Choose UofU_Analog_Parts for the library, then select nmos. The Add Instance
dialog will open. Set the Width to 3.6u (u indicates microns). Click in the schematic editor window
to drop the transistor. You can click a second time to place another transistor. Return to the
Component Browser window and choose pmos. Drop two pMOS transistors. Then return to the
browser and get a gnd and a vdd symbol. When you are in a mode in the editor, you can press ctrl-c
or Esc to get out of it. Other extremely useful commands include Edit Move, Edit Copy, Edit
Undo, and Edit Delete. Edit Properties Object is also useful to change things like transistor
sizes or wire names. Move the elements around until they are in attractive locations. I like to keep
series transistors onew grid unit apart and place pMOS transistors two grid units above the
nMOS. Look at the bottom of the schematic editor window to see what mode you are in. Next,
use Add Pin to create some pins. In the Add Pin dialog, enter a and b. Make sure the direction is
input. The tools are case-sensitive, so

use lower case everywhere. Place the pins, being sure that a is the bottom one. Although pin order
doesnt matter logically, it does matter physically and electrically, so you will get errors if you
reverse the order. Then place an output pin y. Now, wire the elements together. Choose Add Wire
(narrow). Click on each component and draw a wire to where it should connect. It is a good idea to
make sure every net (wire) in a design has a name. Otherwise, youll have a tough time tracking
down a problem later on one of the unnamed nets. Every net in your schematic is connected to a
named pin or to power or ground except the net between the two series nMOS transistors. Choose
Add Wire name Enter mid or something like that as the name, and click on the wire to name it.
Choose Design Check and Save to save your schematic. Youll probably get one warning about a
solder dot on crossover at the
4-way junction on the output node. This is annoying because such 4-way junctions are normal and
common. Choose Check Rules Setup and click on the Physical tab in the dialog. Change Solder
On CrossOver from warning to ignored and close the dialog. Then Check and Save again and the
warning should be gone. If you have any other warnings, fix them. A common mistake is wires
that look like they might touch but dont actually connect. Delete the wire and redraw it. Poke
around the menus and familiarize yourself with the other capabilities of the schematic editor.
V. Logic Verification
Cells are commonly described at three levels of abstraction. The register-transfer level
(RTL) description is a Verilog or VHDL file specifying the behavior of the cell in terms of
registers and combinational logic. It often serves as the specification of what the chip
should do. The schematic illustrates how the cell is composed from transistors or other cells. The
layout shows how the transistors or cells are physically arranged. Logic verification involves
proving that the cells perform the correct function. One way to do this is to simulate the cell and
apply a set of
1s and 0s called test vectors to the inputs, then check that the outputs match expectation. Typically,
logic verification is done first on the RTL to check that the specification is correct. A testbench
written in Verilog or VHDL automates the process of applying and checking all of the vectors. The
same test vectors are then applied to the schematic to check that the schematic matches the RTL.
Later, we will use a layout-versus schematic (LVS) tool to check that the layout matches the
schematic (and, by inference, the RTL). You will begin by simulating an RTL description of the
NAND gate to become familiar with reading RTL and understanding a testbench. In this tutorial,
the RTL and testbench are written in System Verilog, which is a 2005 update to the popular
Verilog hardware description language.
There are many Verilog simulators on the market, including NC-Verilog from Cadence,
VCS from Synopsys, and ModelSim from Mentor Graphics. This tutorial describes how to use NC
Verilog because it integrates gracefully with the other Cadence tools. NCVerilog compiles your
Verilog into an executable program and runs it directly, making it much faster than the
older interpreted simulators. Make a new directory for simulation (e.g. nand2sim). Copy nand2.sv,
nand2.tv,
and testfixture.verilog from the course directory into your new

directory.

mkdir nand2sim
cd nand2sim
cp /courses/e158/10/nand2.sv .
cp /courses/e158/10/nand2.tv .
cp /courses/e158/10/nand2.testfixture testfixture.verilog
nand2.sv is the SystemVerilog RTL file, which includes a behavioral description of a nand2
module and a simple self-checking testbench that includes
testfixture.verilog. testfixture.verilog reads in testvectors
from
nand2.tv and applies them to pins of the nand2 module. After each cycle it compares the output of the
nand2 module to the expected output, and prints an error if they do not match. Look over each of
these files and understand how they work. First, you will simulate the nand2 RTL to practice the
process and ensure that the testbench works. Later, you will replace the behavioral nand2 module
with one generated from your Electric schematic and will resimulate to check that your schematic
performs the correct function. At the command line, type sim-nc nand2.sv to invoke the simulator.
You should see some messages ending with
ncsim> run
Completed 4 tests with 0 errors.
Simulation stopped via $stop(1) at time 81 NS + 0
Youll be left at the ncsim command prompt. Type quit to finish the simulation. If the simulation
hadnt run correctly, it would be helpful to be able to view the results. NC-Verilog has a graphical
user interface called SimVision. The GUI takes a few seconds to load, so you may prefer to run it
only when you need to debug. To rerun the simulation with the GUI, type sim-ncg nand2.sv A
Console and Design Browser window will pop up. In the browser, click on the + symbol beside the
testbench to expand, then click on dut. The three signals, a, b, and y,
will appear in the pane to the right. Select all three, then right-click and choose Send to Waveform
Window. In the Waveform Window, choose Simulation Run. Youll see the waveforms of your
simulation; inspect them to ensure they are correct. The 0 errors message should also appear in the
console. If you needed to change something in your code or testbench or test vectors, or wanted to
add other signals, do so and then Simulation Reinvoke Simulator to recompile everything and bring
you back to the start. Then choose Run again. Make a habit of looking at the messages in the
console window and learning what is normal. Warnings and errors should be taken seriously; they
usually indicate real problems that will catch you later if you dont fix them.
VI. Schematic Simulation
Next, you will verify your schematic by generating a Verilog deck and pasting it into the RTL
Verilog file. While viewing your schematic, click on Tools Simulation NCVerilog to open a
window for the Verilog environment. Note the run directory (e.g. nand2_run1), and press the button
in the upper left to initialize the design. Then press the next button to generate a netlist. Look in the
icfb

window for errors and correct them if necessary. You should see that the pmos, nmos, and nand2
cells were all netlisted. In your Linux terminal window, cd into the directory that was created.
Youll find quite a few files. The most important are verilog.inpfiles,
testfixture.template, and testfixture.verilog. Each cell is netlisted into a different directory under
ihnl. verilog.inpfiles states where they are. Take a look at the netlist and other files.
testfixture.template is the top level module that instantiates the device under test and invokes the
testfixture.verilog. Copy your from your nand2sim directory to your nand2_run1 directory using a
command such as
cp ../nand2sim/testfixture.verilog .
cp ../nand2sim/nand2.tv .
Back in the Virtuoso Verilog Environment window, you may wish to choose Setup Record
Signals. Click on the All button to record signals at all levels of the hierarchy. (This isnt
important for the nand with only one level of hierarchy, but will be helpful later.) Then
choose Setup Simulation. Change the Simulation Log File to indicate simout.tmp sv. This will
print the results in simout.tmp. The sv flag indicates that the simulator should accept
SystemVerilog syntax used in the testfixture.verilog. Set the Simulator mode to Batch and click
on the Simulate button. You should get a message that the batch simulation succeeded. This doesnt
mean that it is correct, merely that it run. In the terminal window, view the simout.tmp file. It
will

give some statistics about

the compilation, then should indicate that the 4 tests were

completed with 0 errors. If the simulation fails, the simout.tmp file will have clues about the
problems. Change the simulator mode to Interactive to rerun with the GUI. Be patient; the GUI
takes several seconds to start and gives no sign of life until then. Add the waveforms again and run
the simulation. You may need to zoom to fit all the waves. For some reason, SimVision doesnt
print the $display message about the simulation succeeding with no errors. You will have to read the
simout.tmp file at the command line to verify that the test vectors passed. If you find any logic errors,
correct the schematic and resimulate.

Silicon Debug
The rapid pace of innovation has created powerful SOC solutions at consumer prices. This
has created a highly competitive market place where billions of dollars can be won by the right
design delivered at the right time. These new designs are produced on processes that
challenge the fundamental laws
to

equipment

of

physics

and

are highly

sensitive

variation. The industry now produces new designs in a

complex world where process and design interaction have created new complex failures that
stand in the way of billion-dollar opportunities. These interactions lead to new types of
defects such as blocked chains, which create noise in the debug/diagnosis process. They also
lead to new types of design issues such as delay defects in combinational and sequential logic.

The challenge is made even greater by the growing complexity in device structure and design
techniques. Multiple design organizations use multiple IP blocks and multiple libraries that need
to work together

throughout

the

process

window,

often

across

multiple

fabs.

These new challenges come at a time when product lifetimes are shrinking, leading to pressure to
reduce time for debug and characterization activities. These problems are seen for the first time at
first silicon. Test the first chips back from fabrication If you are lucky, they work the first time If
not . . . Logic bugs vs. electrical failures Most chip failures are logic bugs from inadequate
simulation or verification Some are electrical failures Crosstalk Dynamic nodes: leakage, charge
sharing Ratio failures A few are tool or methodology failures (e.g. DRC) Fix the bugs and fabricate a
corrected chip Silicon debug (or bringup) is primarily a Non-Recurring Engineering (NRE)
cost (like design) Contrast this with manufacturing test which has to be applied to every part shipped.
Manufacturing test:
A speck of dust on a wafer is sufficient to kill chip Yield of any chip is < 100% Must test
chips after manufacturing before delivery to customers to only ship good parts Manufacturing
testers are very expensive Minimize time on tester Careful selection of test vectors. A test for a
defect will produce an output response which is different from the output when there is no defect
Test quality is high if the set of tests will detect a very high fraction of possible defects Defect level
is the percentage of bad parts shipped to customers Yield is the percentage of defect-free chips
manufactured
Fault Models
Fault models:
Numerous possible physical failures (what we are testing for) Can reduce the number of
failure types by considering the effects of physical failures on the logic functional blocks: called a
Assume that defects will cause the circuit to behave as if lines were stuck at logic 0 or 1 Most
commercial tools for test are based on the stuck-at model Other fault models Stuck open model
for charge retained on a CMOS node Recent use of the transition fault model in an attempt to deal
with delays Path delay fault model would be better for small delay defects, but the large number
of possible paths is an impediment to the use of this fault model.
Test for faults in a circuit:
Approach to generating tests for defects is to map defects to (higher level) faults: develop
fault model, then generate tests for the faults Typical: gate-level stuck-at fault model As
technology shrinks, other faults: bridging faults, delay faults, crosstalk faults, etc. An interesting
point: what is important is how well the tests generated (based on the fault model) will detect
realistic defects The accuracy of the fault model is secondary

Observability and controllability:

Observability: ease of observing a value on a node by monitoring external output pins of the chip
Controllability: ease of forcing a node to 0 or 1 by driving input pins of the chip Combinational logic
is usually easier to observe and control Still, NP-complete problem Finite state machines can be very
difficult, requiring many cycles to enter desired state Especially if state transition diagram is
not known to the test engineer, or is too large
Fault simulation:
Identify faults detected by a sequence of tests Provide a numerical value of coverage (ratio
of detected faults to total faults) Correlation between high fault coverage and low defect level
Faults considered Generally, gate level stuck-at faults Can also evaluate coverage of switch level
faults Can include timing and dynamic effects of failures Although fault simulation takes polynomial
time in the number of gates, it can still be prohibitive for large designs. Static timing analysis
(Primetime, for example) only finds structural long paths
False Path problem:
In order to allow a signal to go through the path, Required Side Inputs: C = 1, A = 1, E = 1 Conflict
due to C = 1 and E = 1 Can use modified test generation algorithms to identify longest true paths in a
circuit CRITIC from UT Primetime+Tetramax from Synopsys.
Boundary scan:
Boundary scan is a method for testing interconnects (wire lines) on print ed circuit boards
or sub-blocks inside an integrated circuit. Boundary scan is also widely used as a debugging method
to watch integrated circuit pin states, measure voltage, or analyze sub-blocks inside an integrated
circuit.
Testing
The boundary scan architecture provides a means to test interconnects and clusters of
logic, memor ies etc. without using physical t est probes. It adds one or more so called 'test cells'
connected to each pin of the device that can selectively override the functionality of that pin.
These cells can be programmed via the JTAG scan chain to drive a signal onto a pin and across an
individual trace on the board. The cell at the destination of the board trace can then be programmed
to read the value at the pin, verifying the board trace properly connects the two pins. If the trace is
shorted to another signal or if the trace has been cut, the correct signal value will not show up at
the destination pin, and the board will be observed to have a fault.

On-Chip Infrastructure
To provide the boundary scan capability, IC vendors add additional logic to each of their
devices, including scan cells for each of the external traces. These cells are then connected together
to form the external boundary scan shift register (BSR), and combined with JTAG TAP (Test
Access Port) controller support comprising four (or sometimes more) additional pins plus control
circuitry.
Some TAP controllers support scan chains between on-chip logical design blocks, with JTAG
instructions which operate on those internal scan chains instead of the BSR. This can allow those
integrated components to be tested as if they were separate chips on a board. On-chip debugging
solutions are heavy users of such internal scan chains.
These designs are part of most Verilog or VHDL libraries. Overhead for this additional
logic is minimal, and generally is well worth the price to enable efficient testing at the board level.
For normal operation, the added boundary scan latch cells are set so that they have no effect
on the circuit, and are therefore effectively invisible. However, when the circuit is set into a test
mode, the latches enable a data stream to be shifted from one latch into the next. Once a complete
data word has been shifted into the circuit under test, it can be latched into place so it drives
external signals. Shifting the word also generally returns the input values from the signals configured
as inputs.
Test Mechanism
As the cells can be used to force data into the board, they can set up test conditions. The
relevant states can then be fed back into the test system by clocking the data word back so that it can
be analyzed.
By adopting this technique, it is possible for a test system to gain test access to a board.
As most of todays boards are very densely populated with components and tracks, it is very difficult
for test systems to physically access the relevant areas of the board to enable them to test the
board. Boundary scan makes access possible without always needing physical probes.
In modern chip and board design, Design For Test is a significant issue, and one common
design artifact is a set of boundary scan test vectors, possibly delivered in S erial Vector For mat
(SVF) or a similar interchange format.
JTAG Test Operations

Devices communicate to the world via a set of input and output pins. By themselves, these
pins provide limited visibility into the workings of the device. However, devices that support
boundary scan contain a shift-register cell for each signal pin of the device. These registers
are connected in a dedicated path around the device's boundary (hence the name). The path
creates a virtual access capability that circumvents the normal inputs and provides direct control of
the device and detailed visibility at its outputs.

[3]

The contents of the boundary scan are usually

described by the manufacturer using a part-specific BSDL file.

Among other things, a BSDL file will describe each digital signal exposed through pin or ball
(depending on the chip packaging) exposed in the boundary scan, as part of its definition of the
Boundary Scan Register (BSR). A description for two balls might look like this:
"541 (bc_1,

*, control, 1)," &

"542 (bc_1,

GPIO51_ATACS1, output3, X, 541, 1, Z)," &

"543 (bc_1,

GPIO51_ATACS1,

"544 (bc_1,

input, X)," &

*, control, 1)," &

"545 (bc_1,

GPIO50_ATACS0, output3, X, 544, 1, Z)," &

"546 (bc_1,

GPIO50_ATACS0,

input, X)," &

That shows two balls on a mid-size chip (the boundary scan includes about 620 such lines, in
a 361-ball BGA package), each of which has three components in the BSR: a control configuring the
ball (as input, output, what drive level, pullups, pulldowns, and so on); one type of output signal; and
one type of input signal.
There are JTAG instructions to SAMPLE the data in that boundary scan register, or
PRELOAD it with values.
During testing, I/O signals enter and leave the chip through the boundary-scan cells. Testing
involves a number of test vectors, each of which drives some signals and then verifies that
the responses are as expected. The boundary-scan cells can be configured to support external
testing for interconnection between chips (EXTEST instruction) or internal testing for logic
within the chip (INTEST instruction).
Board Test Infrastructure
Typically high-end commercial JTAG testing systems allow the import of design 'netlists' from
CAD/EDA systems plus the BSDL models of boundary scan/JTAG complaint devices to
automatically generate test applications. Common types of test include

Scan-path 'infrastructure' or integrity

Boundary-scan device pin to boundary-scan device pin 'interconnect'

Boundary-scan pin to memory device or device cluster (SRAM, DRAM, DDR etc)

Arbitrary logic cluster testing

When used during manufacturing, such systems also support non-test but affiliated applications
such as in-system programming of various types of flash memory: NOR, NAND, and serial (I2C or
SPI).
Such commercial systems are used by board test professionals and will often cost several
thousand dollars for a fully-fledged system. They can include diagnostic options to accurately pinpoint faults such as open circuits and shorts and may also offer schematic or layout viewers to
depict the fault in a graphical manner. Tests developed with such tools are frequently combined
with other test systems such as in-circuit testers (ICTs) or functional board test systems.

UNIT V SPECIFICATION USING VERILOG HDL

Basic concepts
Verilog HDL:
Verilog is a HARDWARE DESCRIPTION LANGUAGE (HDL). A hardware
description Language is a language used to describe a digital system, for example, a
microprocessor or a memory or simple flip-flop. This just means that by using a HDL one
can describe any hardware (digital) at any level.
White space:
White space characters, such as SPACE and TAB and blank lines are ignored by
verilog compiler. Multiple statements are written on a single line such as,
f = a0; if (s == 0) f = a1;
Placing each statement on a separate line and using indentation within blocks of
code, such as an if-else statement are good ways to increase the readability of code.
Documentation in verilog code:
Documentation can be included in verilog code by writing a comment. A short
comment begins with double slash, //, and continues to the end of the line. A long comment
can span multiple lines and is contained inside the delimiters /* and */.
//this is a short comment
/* this is a long comment
That spans two lines */
Operators:
Verilog has operators of three types. They are unary operators which precede the
operand, binary operators which appear between two operands and ternary operators that
have two separate operators that separate three operands.
Example:
x = + y; //+ is unary operator, y is the operand
x = y && z; // && is a binary operator, y and z are operands
x = a ? b : c; // ? : is a ternary operator, a, b, c are operands

Signal values, Numbers:

Verilog supports scalar nets and variables that can be represent individual signals
and vectors that correspond to multiple signals. Each individual signal can have four
possible
values:
0 = logic value 0;
1 = logic value 1;
z = high impedance;
x = unknown value;
The value of a vector variable is specified by giving a constant of the form
[size][ radix]constant
where size is the number of bits in the constant and resix is the number base.
Radices are:
d = decimal;
b = binary;
h = hexadecimal;
o = octal;
when no radix is specified the default is decimal.
Constants are:
0

number 0;

10

decimal number 10;

b10

binary number 10 = (2)10;

h10

hex number 10 = (16)10;

4b100

binary number 0100 = (4)10;

4bx

unknown 4 bit value xxxx;

8hfx

equivalent to 8b1111_xxxx;

Parameters:
A parameter in verilog associates an identifier name with a constant.
Declaration:
parameter n = 4;
parameter s0 = 2b00, s1 = 2b 01; s2 =2b11; s3 = 2b 10;
Nets:
A net represents a node in a circuit. There are two different types of nets are used in
verilog. They are wire and tri.
Wire type net can be employed to connect an output of one logic element in a
circuit to an input of another logic element. Examples of scalar wire declarations:
wire a;
wire a, b;
Example of vector wire declaration:

wire [2:0] s;
wire [1:3] x;
The square brackets are the syntax for specifying a vectors range. The range [R a:Rb]
can be either increasing or decreasing. The indices R a and Rb can be either positive or
negative integers.
The net s can be used as a three bit quantity or each bit can be referred to
individually as s[2], s[1] and s[0]. If a value assigned to s such as s = 3b010, the result
is s[2] = 0, s[2] = 1 and s[0] = 0.
The tri type denotes circuit nodes that are connected in a tri state
fashion. Example:
tri x;
tri [7:0] out;
These nets are treated in the same manner as the wire type and they are used only to
enhance the readability of code includes tri state gates.
Variables:
A variable can be assigned a value in one verilog statement and it retains this value
until it is overwritten in a subsequent assignment statement. There are two types of
variables,
reg and integer.
reg [2:0] out;
The above declarations show that out has three bits. The keyword reg does not
denote a storage element or register.

In verilog code reg variables can be used to

model either combinational or sequential parts of a circuit.

integer x;
Integer variables are useful for describing the behaviour of a module, but they do not
directly correspond to nodes in a circuit.
Memories:
Memory is a two dimensional array of bits. Verilog allows such a structure to be
declared as a variable (reg or integer) that is an array of vectors, such as
reg[3:0] x [1:0]
This statement defines x as two four-bit variables named x[1] and x[0]. Memories
cannot be net types and they cannot be used as ports on a module.

Identifiers
Identifiers are the names of variables and other elements in verilog code.
Rules:
a) Any letter or digit may be used

b) The _underscore and $ characters are used

c) Identifiers must not begin with a digit and it should not be a verilog keyword
Example: legal identifiers are a, x1, a_b and byte.
Illegal identifiers are 1a, +b, a*b and 258.
Verilog is case sensitive, hence k is not the same as K and BYTE is not the same as Byte.
Verilog allows a second form of identifier called an escaped identifier.
These identifiers are begin with (\) backslash character, which can then be followed by any
printable ASCII characters except white spaces.
Example: \234, \sig-name, \x+y
Escaped identifiers should not be used in normal verilog code, they are intended for
use in code produced automatically when other languages are translated into verilog.
Structural level modeling:
In gate level modeling or structural level modeling, the circuit is described in terms of
logic gates. Example: and, or, nand, nor, etc...

The logic gates have one scalar output and

multiple scalar inputs. The first terminal in the list of gate terminals is an output and the other
terminals are
inputs.
a. Gate Primitives
Verilog defines some basic logic gates as part of the language. Gate
primitives used in verilog codes are NOT, AND, OR, NAND, NOR, XOR, XNOR gate. Verilog
includes predefined modules that implement basic logic gates. These gates allow a circuits
structure to be described using gate instantiation statements of the form:
gate_name [instance_name] (output_port, input_port {, input_port});
Here gate_name specifies the gate type and instance_name is an identifier, also it is optional.
Each gate may have different number of ports, the output port listed first, followed
by a variable number of input ports.
Example:

and And1 (output, a, b);

or Or1 (output, a, b);

nor Nor1 (output, a, b);

xnor Xnor1 (output, a, b);

xor Xor1 (output, a, b);

not Not1 (output, input);

nand Nand (output, a, b);

Structural specification of a full adder:

module fulladd (a,b,cin, sum,carry) ;

input a, b, cin;
output sum, carry;
wire p, q, r;
and And1 (p, a, b);
and And2 (q, a, cin);
and And3 (r, b, cin);
or Or1 (carry, p, q, r);
xor Xor1 (sum, a, b, cin);
endmodule

Simplified version of full adder:

module fulladd (a,b,cin, sum,carry) ;

input a, b, cin;
output sum, carry;
wire p, q, r;
and (p, a, b);
and (q, a, cin);
and (r, b, cin);
or (carry, p, q, r);
xor (sum, a, b, cin);
endmodule

Gate delays
A delay, as used in Verilog, is a number of time units it takes to receive a response of
a circuit. In a simple forward combinational circuit this is a time it takes to obtain a change
on an output when an input is altered. Let's say a synchronous circuit is triggered by a
positive clock transition; `the delay' is then the time it takes for the circuit to change its
outputs from the time the clock line is set high rather then a time it takes to change the
outputs since data lines altered.
Types of delays
There are three basic types of delay which can be used:

the time it takes to set an output high (t rise); this applies to a transition which may start from
any (`0', `1', `X' or `Z') state

the time it takes to set an output low (t fall); similarly, applies to transitions which begin in any
state.

and finaly, the time it takes to cut-off a buffer (toff).

Syntax gate with delay:

gate_type #(t_rise, t_fall, t_off) instance_name (output_port,input_port1, input_port2,...);

For example:
and #(1, 3) g1 (o1, i1, i2);
nor #(2) g2 (o2, i3, i4);
The above parameters are specified in sequence and if a cut-off time is to be specified the fall time
cannot be omitted. One more interesting transition is a move of an input from a defined logic state i.e.
either `0' or `1' to an `X' (don't care). In such case, the shortest of the two times t fall, trise is used. For
an inverse case (`X'->`0'|`1') the longer delay is used.

Example: Single bit full-adder

The design below is for a full adder, written using gate-level modelling
techniques. A gate-level model of a 1-bit full-adder:
module full_adder(sum, c_out, a, b, c_in);
output c_out, sum;
input a, b, c_in;
wire p, q, r;
xor #(3,2) x0 (p, a, b);
xor #(3,2) x1 (sum, p, c_in);
and #(2,4) a0 (q, a, b);
and #(2,4) a1 (r, p, c_in);
or #(3) o0 (c_out, q, r);
endmodule

Full Adder: Output - no delays

Full Adder: Output - delays as specified

Operators
Verilog has large number of operators.
Binary Arithmetic Operators
Binary arithmetic operators operate on two operands. Register and net
(wire) operands are treated as unsigned. However, real and integer operands may be
signed. If any bit is unknown ('x') then result is unknown.
Operator Name
+

Addition

Subtraction

Multiplication

Division

Modulus

Relational Operators

Relational operators compare two operands and return a logical value, i. e., TRUE(1)
or FALSE(0). If any bit is unknown, the relation is ambiguous and the result is unknown.
Operator

Name

>

Greater than

>=

Greater than or equal

<

Less than

<=

Less than or equal

==

Logical equality

!=

Logical inequality

Logical Operators
Logical operators operate on logical operands and return a logical value, i. e., TRUE(1)
or FALSE(0). Used typically in if and while statements. Do not confuse logical operators with
the bitwise Boolean operators. For example , ! is a logical NOT and ~ is a bitwise NOT. The first
negates,
e. g., !(5 == 6) is TRUE. The second complements the bits, e. g., ~{1,0,1,1} is 0100.
Operator Name
!

Logical negation

&&

Logical AND

||

Logical OR

Bitwise Operators
Bitwise operators operate on the bits of the operand or operands. For example, the result of A
& B is the AND of each corresponding bit of A with B. Operating on an unknown (x) bit results in
the expected value. For example, the AND of an x with a FALSE is an x. The OR of an x with a
TRUE is a TRUE.
Operator

Name

Bitwise negation

&

Bitwise AND

Bitwise OR

Bitwise XOR

~&
~|
~^ or ^~

Bitwise NAND
Bitwise NOR
Bitwise NOT XOR

Unary Reduction Operators

Unary reduction operators produce a single bit result from applying the operator to all of the
bits of the operand. For example, &A will AND all the bits of A.
Operator Name
&

AND reduction

OR reduction

XOR reduction

~&

NAND reduction

~|

NOR reduction

~^

XNOR reduction

Unary Arithmetic Operators

Operator

Name

Unary Minus

Comments
Changes sign of its operand.

Timing controls
A timing control is either a delay control or an event control.
Delay control:
A delay control delays an assignment by a specified amount of time. A timescale
compiler directive is used to specify the units of time followed by the precision used to
calculate time expressions,
`timescale 1ns/10ps // Units of time are ns. Round times to 10 ps.

Time units may only be s , ns , ps , or fs and the multiplier must be 1, 10, or 100. We can
delay an assignment in two different ways:

Sample the RHS immediately and then delay the assignment to the LHS.

Wait for a specified time and then assign the value of the LHS to the RHS.

Here is an example of the first alternative (an intra-assignment delay):

a = #1 b; // intra-assignment delay
The second alternative is delayed assignment:
#1 a = b; // delayed assignment
These two alternatives are not the same. The intra-assignment delay is equivalent to
the following code:
begin // Equivalent to intra-assignment delay.
hold = b; // Sample and hold y immediately.
#1; // Delay.
a = hold; // Assignment to a. Overall same as a = #1 b.
end
In contrast, the delayed assignment is equivalent to a delay followed by an assignment as follows:
begin // Equivalent to delayed assignment.
#1; // Delay.
a = b; // Assign y to x. Overall same as #1 a = b.
end
Event control:
An event control, delays an assignment until a specified event occurs. Here is the formal
definition:
event_control ::= @ event_identifier | @ (event_expression)
event_expression ::= expression | event_identifier
| posedge expression | negedge expression
| event_expression or event_expression
A positive edge (denoted by the keyword posedge ) is a transition from '0' to '1' or 'x' , or a
transition from 'x' to '1 '. A negative edge ( negedge ) is a transition from '1' to '0' or 'x' , or a
transition from 'x' to '0'. Transitions to or from 'z' do not count. Here are examples of event controls:
module delay_controls; reg X, Y, Clk, Dummy;
always #1 Dummy=!Dummy; // Dummy clock, just for graphics.
// Examples of delay controls:

always begin #25 X=1;#10 X=0;#5; end

// An event control:
always @(posedge Clk) Y=X; // Wait for +ve clock edge.
always #10 Clk = !Clk; // The real clock.
initial begin Clk = 0;
$display("T Clk X Y");
$monitor("%2g",$time,,,Clk,,,,X,,Y);
$dumpvars;#100 $finish; end
endmodule
T Clk X Y
0 0 xx
10 1 x x
20 0 x x
25 0 1 x
30 1 1 1
35 1 0 1
40 0 0 1
50 1 0 0
60 0 0 0
65 0 1 0
70 1 1 1
75 1 0 1
80 0 0 1
90 1 0 0
The dummy clock in delay_controls helps in the graphical waveform display of the results (it
provides a one-time-tick timing grid when we zoom in, for example). Figure shows the graphical
output (white is used to represent the initial unknown values). The assignment statements to 'X'
in the always statement repeat (every 25 + 10 + 5 = 40 time ticks).

FIGURE: Output

from

the

module

delay_controls .

Events can be declared (as named events), triggered, and detected as follows:
module show_event;
reg clock;
event event_1, event_2; // Declare two named events.

always @(posedge clock) -> event_1; // Trigger event_1.

always @ event_1
begin $display("Strike 1!!"); -> event_2; end // Trigger event_2.
always @ event_2 begin $display("Strike 2!!");
$finish; end // Stop on detection of event_2.
always #10 clock = ~ clock; // We need a clock.
initial clock = 0;
endmodule

Procedural assignments conditional statements

Verilog provides procedural statements. Procedural statements are executed in which
order they appear in the code. Verilog syntax requires that procedural statements be
contained inside an always block.
Always blocks:
An always block is a construct that contains one or more procedural statements. It
has the form:

always @ (sensitivity_list)
[begin]
[procedural assignment statements]
[if-else statements]
[case statements]
[while, repeat, and for loops]
[task and function calls]
[end]

Verilog includes several types of procedural statements. When multiple statements

are included in an always block, the begin and end keywords are used, otherwise it can be
omitted.
The sensitivity_list is a list of signals that directly affect the output r esults
generated by the always block.
Example for always block:

always @ (a or b)
begin
sum = a ^ b;
carry = a & b;
end
Since the output variables sum and carry depend on a and b, these signals are
included in the sensitivity list, separated by a keyword or.

If the value of a signal in the

sensitivity list changes, then the statements inside the always block are evaluated in the order
presented.
Procedural assignment statements:
Any signal assigned a value inside an always block has to be a variable of type reg or
integer. A value is assigned to a variable with a procedural assignment statement. There are
two types of assignments:
a) blocking assignments
b) non-blocking assignments
Blocking assignments are denoted by the symbol (=).
The term blocking means that the assignment statement completes and
updates its left hand side before the subsequent statement is evaluated.
Example:
sum = a + b;
x = sum [0];
The first statement sets sum using the current values of a and b, and then the second
statement sets p according to this new value of sum.
Verilog also provides non-blocking assignments and it can be denoted by (<=).
Example:
s <= a + b;
p <= s [0];
At simulation time ti, the statements are still evaluated in order, but they both use the values
of variables that exist at the start of the simulation time, t i.
The first statement determines a new value for s based on the current values of a and b, but
s is not actually changed to this value until all statements in the associated always block have
been evaluated. So, the value of p at time ti is based on the value of s at time ti-1.
Summary:
For blocking assignments, the values of variables seen at time t i by each statement are the
new values set in ti by any preceding statements in the always block.

For non-blocking assignments, the values of variables seen at time t i are the values set in time
ti-1.
The if-else statement:
General form of the if-else statement is given in figure.

if (expression 1)
begin
statement;
en
d
else if (expression 2)
begin
statement;
e
nd
els
e
beg
in
statement;
en
d

Figure
If expression 1 is true, then the first statement is evaluated. When multiple
statements are involved, they have to be included inside a begin-end block.

Example:
Verilog code for 2 x 1 multiplexer

always @ (a or b or s)
if (s == 0)
f = a;
else
f = b;

The code defines a 2 x 1 multiplexer with data inputs a and b, with select input s and output f.
Case statement:
General form of case statement is shown in the figure.
case (expression)
alternative1: begin
statement:
end
alternative2 : begin
statement;
end
[default: begin
statement;
end]
endcase

The bits in the expression are called controlling expression, it checked for a match with each
alternative. The first successive match causes the associated statements to be evaluated. Each digit
in each alternative is compared for a correct match of the four values 0, 1, x and z.
A special case is the default clause, which takes effort if no other alternative matches.
Example:
Verilog code for 2 x 1 multiplexer.

always @ (x or y or s)
case (s)
1b0: f = x;
1b1: f = y;
endcase

In this example, s can have four values 0, 1, x, z. If s = 0 means output f = x, and if s = 1 means output
f = y.
Example: Verilog code for full adder using case statement.

module fulladd (a, b, cin, sum, carry);

input a, b, cin;
output sum, carry;
reg sum, carry;
always @ (a or b or cin)
begin
case ({a, b, cin})
3b000: {carry, sum} = b00;

3b001: {carry, sum} = b01;

3b010: {carry, sum} = b10;
3b011: {carry, sum} = b01;
3b100: {carry, sum} = b10;
3b101: {carry, sum} = b10;
3b110: {carry, sum} = b11;
3b111: {carry, sum} = b00;
endcase
end
endmodule

Loop statements:
Verilog code includes four types of loop statements: for loop, while loop, repeat loop and
forever loop. Synthesis tool support the for loop.
General form of for loop is given below.

for (initial_index; terminal_index; increment)

begin
statement;
end
The initial_index is evaluated once, before the first loop iteration and typically performs the
initialization of the integer loop control variable. In each loop iteration, the begin-end block
is performed. The terminal_index condition is checked and if it is true, then another loop iteration
is done.
Example:
Verilog code for bit-counting using for loop.

always @ (x)
begin
count=0;
for (x = 0; x < n; x = x + 1)
count = count + b (x);
end

General form of while loop and repeat loop is given below.

while (condition)
begin
statement;
end

repeat (constant_value)
begin
statement;
end

Tasks and Functions

Tasks are like procedures in other programming languages, e. g., tasks may have zero or
more arguments and do not return a value. Functions act like function subprograms in other
languages. Except:
1. A Verilog function must execute during one simulation time unit. That is, no time
controlling statements, i. e., no delay control (#), no event control (@) or wait statements, allowed.
A task can contain time controlled statements.
2. A Verilog function cannot invoke (call, enable) a task; whereas a task may call other tasks
and functions.
The definition of a task is the following:
task <task name>; // Notice: no list inside ()s
<argument ports>
<declarations>
<statements>
endtask
An invocation of a task is of the following form:
<name of task> (<port list>);
where <port list> is a list of expressions which correspond to the <argument ports> of the
definition. Port arguments in the definition may be input, inout or output. Since the <argument
ports> in the task definition look like declarations, the programmer must be careful in adding
declares at the beginning of a task.

Testing tasks and functions

module tasks;
task add;

// task definition

input x, y; // two input argument ports

output z;

// one output argument port

reg R;

// register declaration

begin
R = 1;
if (x == y)
z = 1 & R;
else
z = 0;
end
endtask
initial begin: init1
reg p;
add(1, 0, p); // invocation of task with 3 arguments
$display("p= %b", p);
end
endmodule
The parameters input and inout are passed by value to the task and output and inout
parameters are passed back to invocation by value on return. Call by reference is not
available.
Allocation of all variables is static. Therefore, a task may call itself but each invocation of
the task uses the same storage, i. e., the local variables are not pushed on a stack. Since concurrent
threads may invoke the same task, the programmer must be aware of the static nature of storage
and avoid unwanted overwriting of shared storage space.
The purpose of a function is to return a value that is to be used in an expression. A function
definition must contain at least one input argument. The passing of arguments in functions is the
same as with tasks (see above). The definition of a function is the following:
function <range or type> <function name>;
<argument ports>
<declarations>
<statements>

endfunction
where <range or type> is the type of the results passed back to the expression where the function
was called. Inside the function, one must assign the function name a value. Below is a function
which is similar to the task above.
Testing functions
module functions;
function [1:1] add2; // function definition
input x, y;
reg R;

// two input argument ports

// register declaration

begin
R = 1;
if (x == y)
add2 = 1 & R;
else
add2 = 0;
end
endfunction
initial begin: init1
reg p;
p = add2(1, 0); // invocation of function with 2 arguments
$display("p= %b", p);
end
endmodule
Data flow modeling
In verilog coding gate-level modeling works well due to the number of gates are
less; if large number of gates are used in a circuit then this type of modeling will be
complicated. Dataflow modeling is a powerful approach to implement large circuit. This
modeling becomes a popular approach as logic synthesis tools have become difficult.
Continuous assignments:
Continuous assignments are one type of concurrent statements. While gate
instantiations allow the description of a circuits structure, continuous assignments allow the
description of a circuits function.
General form of the continuous statement is
assign net_assignment {, net_assignment};

the net_assignment can be any expression involving the operators.

Example:
assign carry = (a & b) | (a & c) | (b & c);
assign sum = a ^ b ^ c;
multiple assignments are specified in one assignment, using commas to separate the
assignment.
Example:
assign carry = (a & b) | (a & c) | (b & c), sum = a ^ b ^ c;
Example for multibit assignment is
wire [1:3] x, y, z;
.
.
.
assign z = x & y;
This results in z1 = x1y1, z2 = x2y2, z3 = x3y3.
Delays in dataflow modeling:
Delay values control the time between the change in a right hand side operand
and when the new value is assigned to the left hand side. Delays in continuous assignment
statements can be specified in three ways:
a) Regular assignment delay
b) Implicit continuous assignment delay
c) Net declaration delay
Regular assignment delay:
In this way, the delay value can be specified after the keyword assign.
Example:
assign #10 c = a & b;
a
b
c xxxxx
time

10

20

30

60 70

80

85

1. When the input a and b go to high at time 20, c goes to high at 10 time units later.

2. When a goes low at 60, c goes to low at 70.

3. When a changes to high at 80, but it goes down to low before 10 time units have over and
done.
4. So, at the time of recomputation, 10 units after 80, a is 0. Thus, c gets the value 0.
Implicit continuous assignment delay:
Implicit continuous assignment method is used to specify both a delay and an assignment
on the net.
Example:
wire c;
assign #10 c = a & b;
In implicit continuous assignment delay we can write the above as:
wire #10 c = a & b;

Net declaration delay:

Net declaration delays can be used in gate level modeling. Delay can be specified on a net
when it is declared without putting a continuous assignment on the net. If a delay is specified on a
net c, then any value change applied to the net c is delayed accordingly.
Example:
wire c;
assign #10 c = a & b;
We can write the above by using net declaration delay as:
wire #10 c;
assign c = a & b;
Switch level modelling:
This is another type of modeling used in verilog HDL. In this modeling, MOS switches are
used. In verilog HDL, transistors are also known as switches that either conduct or are open.

MOS Switches
Models of MOS networks consist following four primitive types:
nmos
pmos
rnmos
rpmos
The pmos keyword stands for PMOS transistor and the nmos keyword stands for NMOS
transistor. PMOS and NMOS transistors have relatively low impedance between their sources and
drains when they conduct.
The rpmos keyword stands for resistive PMOS transistor and the rnmos keyword stands for
resistive NMOS transistor. Resistive PMOS and resistive NMOS transistors have significantly higher
impedance between their sources and drains when they conduct than PMOS and NMOS
transistors have. These four gate types are unidirectional channels for data similar to the bufif gates.
Declarations of these gates begin with one of the following keywords:
pmos
nmos
rpmos
rnmos
The delay specification follows the keyword. A terminal list completes the declaration.
The delay specification can be 0, 1, 2, or 3 delays. If there is no delay, there is no delay through the
switch. A single delay determines the delay of all output transitions. If the specification contains 2
delays, the first delay determines the rise delay, the second delay determines the fall delay, and the
smaller of t he
2 delays specifies the delay of transitions to Z and X. If there are 3 delays, the first delay specifies
the rise delay, the second delay specifies the fall delay, the third delay determines the delay of
transitions to Z, and the smallest of the three delays applies to transitions to X. Delays on transitions
to H and L are the same as delays on transitions to X.
These four switches have one output, one data input, and one control input. The first terminal
in the terminal list connects to the output, the second terminal connects to the data input, and the third
terminal connects to the control input.
Some combinations of data input values and control input values cause these switches
to output either of two values, without a preference for either value. These switches logic tables
include
2 symbols representing such unknown results. The symbol L represents a result which has a value of
0
or Z. The symbol H represents a result which has a value of 1 or Z.
pmos
rpmos

CONTROL
01xz

pmos
rpmos

CONTROL
01xz

D0
A1
Tx
Az

0zLL
1zHH
xzxx
zzzz

D0
A1
Tx
Az

0zLL
1zHH
xzxx
zzzz

Logic tables for pmos, rpmos, nmos, and rnmos gates

The following example declares a pmos switch:
pmos (out, data, control);
The output is out, the data input is data, and the control input is control.
Logic symbol:
Nmos switch:
Data

out
Control

Pmos switch:
Data

out

Control

Bidirectional Pass Switches

Declarations of bidirectional switches begin with one of the following keywords:
tran
tranif1
tranif0
rtran
rtranif1
rtranif0
Delay specification follows the keywords in declarations of tranif1, tranif0, rtranif1, and
rtranif0; the next item is the optional identifier. A terminal list completes the declaration.
The delay specifications for tranif1, tranif0, rtranif1, and rtranif0 devices can be 0, 1, or
2 delays. If there is no delay, the device has no turn-on or turn-off delay. If the specification
contains one delay, that delay determines both turn-on and turn-off delays. If
there are 2 delays, the first delay specifies the turn-on delay, and the second delay specifies the turnoff delay.

These six devices do not delay signals propagating through them. When these devices are
turned off they block signals, and when they are turned on they pass signals.
The tranif1, tranif0, rtranif1, and rtranif0 devices have three items in their terminal lists.
Two are bidirectional terminals that conduct signals to and from the devices, and the other
terminal connects to a control input. The terminals connected to inouts precede the
terminal connected to the control input in the terminal list.
The tran and rtran devices have terminal lists containing two bidirectional terminals.
The bidirectional terminals of all six of these devices connect only to scalar nets or bit-selects
of expanded vector nets.
The following example declares a tranif1:
tranif1 (inout1, inout2, control);
The bidirectional terminals are inout1 and inout2. The control input is control.
CMOS Gates
The cmos gate is the combination of a pmos gate and an nmos gate. The rcmos (resistive
cmos) gate is the combination of an rpmos gate and an rnmos gate. The combined gates in
these configurations share data input and data output terminals, but they have separate control inputs.
Cmos switch:
pcontrol

data

out
ncontrol
Declarations of these gates begin with one of these keywords:
cmos
rcmos
The delay specification can be 0, 1, 2 or 3 delays. If there is no delay, there is no delay

through the gate. A single delay specifies the delay for all transitions. If the specification contains
two delays, the first delay determines the rise delay, the second delay determines the fall delay,
and the smaller of the two delays is the delay of transitions to Z
and X. If the specification contains three delays, the first delay controls rise delays, the second
delay controls fall delays, the third delay controls transitions to Z, and the smallest of the three
delays applies to transitions to X. Delays in transitions to H or L are the same as delays in transitions
to X.
The cmos and rcmos gates have a data input, a data output and 2 control inputs. In the
terminal list, the first terminal connects to the data output, the second connects to the data input, the
third connects to the n-channel control input and the last connects to the p-channel

control input.
The equivalence of the cmos gate to the pairing of an nmos gate and a pmos gate is detailed
in the following explanation:
cmos (w, datain, ncontrol, pcontrol);
is equivalent to:
nmos (w, datain, ncontrol);
pmos (w, datain, pcontrol);
pullup and pulldown Sources
Declarations of these sources begin with one of the following keywords:
pullup
pulldown
A strength specification follows the keyword and an optional identifier follows the strength
specification. A terminal list completes the declaration.
A pullup source places a logic value of one on the nets listed in its terminal list. A pulldown
source places a logic value of zero on the nets listed in its terminal list. The signals that these
sources place on nets have pull strength in the absence of a strength specification. There are
no delay specifications for these sources because the signals they place on nets continue throughout
simulation without variation.
The following example declares two pullup instances:
pullup (strong1, strong0)(neta),(netb);
In this example, one gate instance drives neta, the other drives netb.
Behavioural modeling
In this section we are going to discuss about the behavioural features of Verilog.
Modules
A circuit or subcircuit can be explained in verilog is called a module. The general structure
of a module declaration is given below:
module module_name [(port_name {, port_name})];
[parameter declarations]
[input declarations]
[output declarations]
[inout declarations]

[wire or tri declarations]

[reg or integer declarations]
[function or task declarations]
[assign continuous assignments]
[initial block]
[always blocks]
[gate instantiations]
[module instantiations]
endmodule
General form of a module
The module has name, module_name, which can be any valid identifier followed by
a list of ports. The name port refers to an input or output connection in an electrical circuit.
The ports used in verilog code are:
a) input port
b) output port
c) inout port
these ports can be either scalar or vector.
Example:
input a,b,c;
input [2:0]a,b;
output s,c;
inout [5:0]q;
output [3:0]s;
wire x,y;
wire [5:0]y;
reg [ 2:0]x;

All procedural statements occur in blocks that are defined inside modules. There are two
kinds of procedural blocks: the initial block and the always block. Within each block, Verilog uses a
begin and end to enclose the statements. Since initial blocks are ignored during synthesis, only
always blocks are discussed. Always blocks usually take the following format:
always
begin
statement
.....
end
where each statement is a procedural assignment line terminated by a semicolon.
Module Declaration
In the module declaration, the I/O ports of the circuit are declared. Each port has a name and a mode
(in, out, and inout) as shown in the example below.
module ex (A, B, C, D, E);
input A, B, C;
output D;
inout E;
wire D, E;
...
assign E = oe ? A : 1'bz;
assign D = B & E;
...
endmodule
The input and output ports defined in the module declaration called ex are the basic input and
output I/O signals for the design. The inout port in Verilog is analogous to a bi-directional I/O pin on
the device with the data flow for output versus input being controlled by the enable signal to the
tristate buffer.
The preceding example describes E as a tristate buffer with a high-true output enable signal.
If oe = 1, the value of signal A will be output on the pin represented by E. If oe = 0, then the buff er
is in high impedance (Z) and any input value driven on the pin E (from the external logic) will
be brought into the device and fed to the signal represented by D.

Verilog Assignments
There are two forms of assignment statements in the Verilog
language:

Continuous Assignments

Procedural Assignments

Continuous Assignments
Continuous assignments are used to model combinatorial logic in a concise way. Both
explicit and implicit continuous assignments are supported. Explicit continuous assignments are
introduced by the assign keyword after the net has been separately declared. Implicit
continuous assignments combine declaration and assignment.
Example of an explicit continuous assignment:
wire x;
.....
assign x = select ? b : a;
Example of an implicit continuous assignment:
wire temp_hold = a | b;
Note Continuous assignments are only allowed on wire and tri data types.
Procedural Assignments
Procedural assignments are used to assign values to variables declared as regs and are introduced by
always blocks, tasks, and functions. Procedural assignments are usually used to model registers and
FSMs.
Different statements can be used in a process:

Variable and signal assignment

If... else statement

Case statement

For loop statement

Function and task call

The following sections provide examples of each of these

statements.

if...else statement
If... else statements use true/false conditions to execute statements. If the expression
evaluates to true, the first statement is executed. A block of multiple statements may be executed
using begin and end keywords. If...else statements may be nested. The following example shows how
a MUX can be described using an If...else statement.
Example MUX Description Using If... Else Statement
module mux4 (sel, p, q, r, s, y);
input [1:0] sel;
input [1:0] p, q, r, s;
output [1:0] y;
reg [1:0] y;
always @(sel or p or q or r or s)
begin
if (sel[1])
if (sel[0])
y = s;
else
y = r;
else
if (sel[0])
y = q;
else
y = p;
end
endmodule
Case statement
Case statements perform a comparison to an expression to evaluate one of a number of parallel
branches. The Case statement evaluates the branches in the order they are written. The first branch
that evaluates to true is executed. If none of the branches match, the default branch is executed.
Note Do no use unbounded integers in case statements. Always bound integers to a specific number
of bits, or results will be unpredictable.

Casez treats all z values in any bit position of the branch alternative as a don't
care.
Casex treats all x and z values in any bit position of the branch alternative as a don't
care.
The question mark (?) can be used as a "don't care" in any of the preceding case statements. The
following example shows how a MUX can be described using a Case statement.
Example MUX Description Using Case Statement
module mux4 (s a, b, c, d, y);
input [1:0] s;
input [1:0] a, b, c, d;
output [1:0] y;
reg [1:0] y;
always @(s or a or b or c or d)
begin
case (s)
2'b00: y = a;
2'b01: y = b;
2'b10: y = c;
default:y = d;
endcase
end
endmodule
The preceding Case statement will evaluate the values of the input s in priority order. To avoid
priority processing, it is recommended that you use a parallel-case Verilog meta comment which
will ensure parallel evaluation of the sel inputs as in the following.
Example:
always @(sel or a or b or c or d) //synthesis parallel_case
For and Repeat loops
When using always blocks, repetitive or bit slice structures can also be described using the "for"
statement or the "repeat" statement.

The "for" statement is supported for:

Constant bounds

Stop test condition using operators <, <=, > or >=

Next step computation falling in one of the following specifications:

o

var = var + step

var = var - step

(where var is the loop variable and step is a constant value).

The repeat statement is only supported for constant values. The following example shows the use of a
For Loop.
Example For Loop Description
module countzeros (a, Count);
input [7:0] a;
output [2:0] Count;
reg [2:0] Count;
reg [2:0] Count_Aux;
integer i;
always @(a)
begin
Count_Aux = 3'b0;
for (i = 0; i < 8; i = i+1)
begin
if (!a[i])
Count_Aux = Count_Aux+1;
end
Count = Count_Aux;
end
endmodule
Blocking Versus Non-Blocking Procedural Assignments

The # and @ time control statements delay execution of the statement following them until the
specified event is evaluated as true. Use of blocking and non-blocking procedural assignments
have time control built into their respective assignment statement.
The # delay is ignored for synthesis.
The syntax for a blocking procedural assignment is shown in the following
example:
reg a;
a = #10 (b | c);
or
if (in1) out = 1'b0;
else out = in2;
As the name implies, these types of assignments block the current process from continuing to
execute additional statements at the same time. These should mainly be used in simulation.
Non-blocking assignments, on the other hand, evaluate the expression when the statement
executes, but allow other statements in the same process to execute as well at the same time.
The variable change only occurs after the specified delay.
The syntax for a non-blocking procedural assignment is as follows:
variable <= @(posedge or negedge bit) expression;
The following shows an example of how to use a non-blocking procedural
assignment. if (in1) out <= 1'b1;
else out <= in2;
Test bench:
Test bench is another verilog code that creates a circuit involving the circuit to be
tested. This code will send different inputs to the code under test and get the output and displays to
check the accuracy.
Example:
The verilog code for AND gate is given below. It is an example for simple testbench.

module

andgate_tb;

wire

t_y;

reg
andgate

t_a,
my_gate(

t_b;

.a(t_a),

.b(t_b),

.y(t_y)

);

initial
begin
$monitor(t_a,

t_b,

t_y);

t_a

1'b0;

t_b

1'b0;

t_a

1'b0;

t_b

1'b1;

t_a

1'b1;

t_b

1'b0;

1'b1;

#5

#5

#5
t_a
t_b = 1'b1;
end
endmodule
Example:
Teset bench for full adder:
module adder_using_always ();
reg a, b;
reg s, c;
always @ (a or b)
begin
{c,s} = a + b;
end
initial begin
$monitor (" A = %b
B = %b CARRY = %b SUM = %b",a,b,c,s);
#10 a = 0; b = 0;

#10 a = 1;
#10 b = 1;
#10 a = 0;
#10 b = 0;
#10 $finish;
end
endmodule

Example:
Test bench for D-latch
module dlatch();
reg q;
reg d, en;
always @ (d or en)
if (en) begin
q = d;
end
initial begin
$monitor (" ENABLE = %b D = %b Q = %b", en, d, q);
#1 en = 0;
#1 d = 1;
#1 en = 1;
#1 d = 0;
#1 d = 1;
#1 d = 0;
#1 en = 0;

#10 $finish;
end
endmodule
Example:
Test bench for D-flip flop
module dff ();
reg clk,reset,preset,d;
reg q;
always @ (posedge clk or posedge rst or posedge preset)
if (reset) begin
q <= 0;
end else if (preset) begin
q <= 1;
end else begin
q <= d;
end
Testbench code here
initial begin
$monitor("CLK = %b RST = %b PRESET = %b D = %b Q =
%b",
clk,rst,preset,d,q);
clk = 0;
#1 rst = 0;
preset = 0;
d

= 0;

#1 rst = 1;
#2 rst = 0;
#2 preset = 1;
#2 preset = 0;
repeat (4) begin
#2 d
end

= ~d;

#2 $finish;
end
always
#1 clk = ~clk;
endmodule