Novel Global Elmore Delay Optimized Model with Improved Elmore Delay Estimation to Reduce Power Consumption and Delay

© 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 1091
Novel Global Elmore Delay Optimized Model with Improved Elmore
Delay Estimation to Reduce Power Consumption and Delay
Arun Kundu1, Manoj Kumar2
Research Scholar, ECE Department1, Om Sterling Global University, Hisar, Haryana (India)2
Associate professor, ECE Department, Om Sterling Global University, Hisar, Haryana (India)2
--------------------------------------------------------------------------***-----------------------------------------------------------------------
Abstract – The most crucial aspects of VLSI interconnects
as technology has advanced have been size and speed. When
developing an integrated circuit, an IC designer must deal
with scaling issues in interconnects, which are referred to as
the fundamental building block that joins two or more
blocks. Interconnect in VLSI circuits has an increasingly
significant impact as scale grows. The chip's key electrical
properties are entirely within its control. Scale-down
technology causes interconnects to vary in size, bringing
them closer together and potentially having an impact on the
characteristics of the circuit. Pulse and Ramp inputs are used
in lumped and circulated connection circuits to reduce
latency and power consumptio., however, we introduce a
novel interrelate arrangement in this study with enhanced
Elmore delay estimates. To control these parameters, a
number of RC structures have already been described. The
suggested model is estimated and theoretically justified. It
has been found that the RC structure's power consumption
and delay are linearly related. When compared to the earlier
Elmore delay calculations, the planned organization with
enhanced Elmore delay estimation indications a delay
enhancement of 64.25% in lumped circuits and 68.75% in
circulated circuit., which contributes to improving the
interconnect circuit's overall speed.
Key Words – Interconnects, Delay, Power Consumption,
Copper, VLSI
1. INTRODUCTION
Interconnects are discussed as primary constituent
elements of ICs presently. These are the metallic bond that
allows electrical connections within effective systems to
convey and allocate signals and power within the circuit.
Advancement in electronic components resulted ICs.
However, as the patch cord feature size shrinks, signal
integrity issues begin to dominate while
speed, performance, area, and cost characteristic improve.
According to the International Technology Roadmap for
Semiconductors (ITRS), future nanoscale circuits will
contain over a billion transistors and operate at speeds
exceeding 10GHz [1]. The Elmore delay estimate is the
fastest estimate with low complexity, but only the
first instant of the impulse response is considered in
the delay calculation. There are also other approaches that
consider higher-order moments for more accurate delay
calculations and do more analytical work. In addition, many
different models other than the Elmore delay model are
considered for fast and accurate computation of circuits.
Here encoding strategy targeted to accomplish a cut down
of energy consumption, enhancing operating speed and
linkage connections for within-chip conductors [2]. The
encoding method cut down count of total alterations hence
cutting down the highest possible cross-communication by
lessening the connecting-cum-shifting action. The cut down
in switching action deducts the dynamic energy
consumption. The different Interconnect technologies
suggested by different researchers to cut down the cross-
communication issue in circuits are considered in [3].
Reviewing complete model progress of Interconnect
designing in VLSI with interconnect portraying different
two- or three-dimensional field analyzers; Interconnect
design bibliotheca creation, and criterion extirpation [4].
Analytical Interconnect designing were discussed too.
Interconnects operation was optimized by plotted devising
of conductor diameter with model process focused on
Interconnect [5]. Here, two uncomplicated conductor
stiffening mechanisms for VLSI augmentation has been
suggested.
Considering different outlines developed a specified fix for
Interconnect stall time figuring portraying with ideal
conductor-resizing, concurring handler conductor- resizing
and cushion conductor resizing [6-7]. The framework is
wholly customary also useful in unraveling empath issues
as one of voltage source lacework path. Established after
relating various design that other methods work very
speedily hence, handful in a planning mechanism for
outcome-oriented model. A 2π-design depicting tight cross-
communication architecture is suggested [8]. Fixed slab
allocation issue again cutting down on cross-
communication with the procedure. Elmore delay design
suggested conductor diameter issue having [9] examined
the impetus of Interconnect speed lag in the time lagging
Elmore design. The abstracts are noted at the juncture of
real model circuit [10]. Generalized procedure for aligned
RLC/RC circuit is represented here [11].
Elmore delay efficiently evaluated outcome of time-lag of
otherwise dull plot. Observed further lag reduced by
employing following methods. Hence, the centered the
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 10 Issue: 06 | Jun 2023 www.irjet.net p-ISSN: 2395-0072

chapter to re-establish the Interconnect circuit to minimize
time-lag with energy dissipation with the Elmore lag
approximation design. Interconnect architecture enhanced
approximation Elmore time-lag has been suggested in
present chapter so as to cut the delay and energy
dissipation. Out of these, energy dissipation is priority as
diminishing Interconnect size lengthwise. In the research,
the research is focused around:
● Conductor dimensions besides the scaled circuit
dimensions.
● Interconnects intermediate gap.
● Permittivity of substance or material used.
● The material specific resistivity.
2. ADVANCED PROCEDURE
Figure 1 shows the RC circuit both lumped and distributed
for lag approximation. The lumped RC circuit has overall
lag ԏ = RC, representing by lumped network. When the step
function is inputted, the point lag for lumped circuit 50% is
0.7RC. On the other hand, for distributed, the lag is figured
by overall conductor R and C. An RC ladder having n
sections of accumulated L length. When a step function is
inputted, the point lag of distributed circuit 50% is 0.4 RC.
Figure 1(a) RC Lumped circuit (b) RC Distributed circuit
In the present chapter, the simplest architecture has been
proposed. In the above arrangement energy dissipation is
varying as function of length. For more precise relation
between energy and length, an architecture having
enhanced Elmore time-lag approximation having pulse and
ramp as input function is suggested. The estimation is done
for both lumped and distributed circuit of the architecture.
The distributed RC circuit is to have improved operation.
The Interconnect circuit is classified as sub-circuits for
more efficient evaluation in terms of power while time-lag
has constant value. Fort both optimized values, an
Interconnect architecture having Elmore time-lag
approximation. The architecture has resistance with
capacitance has two sections with half the values. This
possesses negligible complexity with ease in simulation put
together in a manner to upside-down having the other not
only modified C value and resistance too. For lumped
architecture, the RC are joint in a manner to give ‘T’.
Figure 2Proposed lumped Interconnect structure
The operational arguments calculated for various lengths
circuits from 1mm to 10mm.
4. RESULTS AND DISCUSSIONS
The Elmore time-lag approximation gives the calculations
having lesser lag to the simulated values, although values
are similar with calculated, the operational arguments
deviations with the lengths.
(a)
0
100
200
300
400
500
600
1 2 3 4 5 6 7 8 9 10
Lumped
Interconnect Length
Lumped
Distributed
Vi
n
Vo
(b
R1 R2 R3
C1 C2
C3
C
(a)
Vin Vout
R
C2
C1
R
C2
R1
R2
C1
3. ADVANCED INTERCONNECT ARCHITECTURE
WITH ENHANCHED ELMORE LAG
APPROXIMATION

(b)
Figure 3For pulse function as input (a) Power vs Length
(b) Delay vs Length
Figure 3 represents relations among Interconnect length
and power, lumped circuits power shoots-up with the
increasing length but distributed circuit power increasing
with length only just. The lengthier wire will dissipate more
power as evident in figure 3. A delay –length plot, lumped
circuit and distributed circuit with different lengths time-
lag is always increasing linearly with Interconnect length.
Also, relationship in the plot has a +ve slope. A strictly
Monotonic behavior for impulse response of time-lag is
verified.
(a)
(b)
Figure 4For ramp function as input (a) Power vs Length
(b) Delay vs Length
Figure 4 (a) and Figure 4 (b) represent power vs length
plot, lumped circuit power is increasing with length but in
distributed power, first increasing with length but later
decreasing with length. The lengthier wire,the more will
consume more power and lesser power is dissipated in
distributed circuit refer to figure 4. The time-lag is not
varying as much with pulse function plot, but energy
dissipation shows great effect. The linear relation shown in
the plot has +ve slope and linearly increasing time-lag
verified the monotonic behavior for impulse function.
CONCLUSION AND FUTURE SCOPE
In the present chapter, Interconnect architecture having
enhanced Elmore time-lag approximation design has been
suggested and studied along with simulation results.
Computations of Elmore lag having lengthier wires with
pulse and ramp functions was studied for lumped as well
distributed segments for various sizes. Finally, arrive at
conclusive for pulse function lumped RC circuit time-lag
and energy consumption are raised simultaneously in
linear way effectively as length increases but in case of
distributed RC, time-lag with length is raised, meanwhile
energy consumption rise is ever so small. And for ramp
function, the time-lag does not rise as with pulse function,
for each of lumped and distributed circuit while energy
dissipation, rises with length in a straight but sloped line
with dissipation rising with length then falling while the
length is raised further. The operational arguments of
distributed circuit have comparatively efficiency higher
than the lumped circuit.
A meliorated Elmore time-lag measurement to cut-down ‘τ’
circuit is studied. Major concerns in interconnect design
were Energy dissipation with time-lag too and needed to be
regulated by modeling or designing an improved RC design.
For each of lumped and distributed RC circuit a linear plot
0
100
200
300
400
500
600
1 2 3 4 5 6 7 8 9 10
Lumped
Interconnect Length
Lumped
Distributed
-5.00E-07
0.00E+00
5.00E-07
1.00E-06
1.50E-06
2.00E-06
2.50E-06
3.00E-06
Lumped
Power Consumption
Lumped
Distributed
-50
0
50
100
150
200
250
300
350
400
450
1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8
Lumped
Delay (ns)
Lumped
Distributed

between energy dissipation as well time-lag has replicated.
For the above sake evaluation economy must be raised
though only just, the architecture outcomes were excellent
in terms of scaled energy dissipation and time-lag within
the chip. Finally, the future scope of this research, various
materials abducted to realize the operational parameter
study of VLSI chip and compared to existing Cu
interconnects. Also, with scaling down of VLSI chips,
inductance becomes considerable in the Interconnect
architecture, and it should also be given due attention.
REFERENCES
[1] Verma, S., K., & Kaushik, B., K. (April 2012). A Bus
Encoding Method for Crosstalk and Power Reduction
in RC Coupled VLSI Interconnects. International
Journal of VLSI Design & Communication Systems
(VLSICS). 3(2). 29-39. DOI:10.5121/vlsic.2012.3203.
[2] Hodges, D., Jackson, H., & Saleh, R. (2004). Analysis
and Design of Digital Integrated Circuits in Deep Sub-
micron Circuits. 3rd ed. Interconnect Design. (pp. 441-
477). New York, NY, USA: McGraw-Hill (Inc.) Higher
Educ.
[3] Sharma, A., & Duggal, D. (April 2016). VLSI
Interconnect Delay Crosstalk Models - A Review.
International Journal of Innovations in Engineering and
Technology (IJIET). 6(4). 2319-1058.
[4] Jung, W., Oh, S., Kong, J. , & Lee, K. (2000). Interconnect
Modeling in Deep-Submicron Design. in IEICE Trans.
Electron. E83-C(8). 1311-1316.
[5] Cong, J. & Pan, Z. (March2002). Wire Width Planning
for Interconnect Performance Optimization. in IEEE
Transactions on Computer-Aided Design of Integrated
Circuits and Systems. 21(3). 319 - 329.
[6] Cong, J., & Pan, D., Z. (Jan1999). Interconnect Delay
Estimation Models for Synthesis and Design Planning.
Proceedings of ASP-DAC 99 Asia & South Pacific Design
Automation Conference.
DOI: 10.1109/ASPDAC.1999.759720.
[7] Vittal, A., & Marek-Sadowska, M. (June 1994). Minimal
Delay Interconnect Design Using Alphabetic Trees.
Paper presented at 31st Design Automation
Conference. IEEE. DOI: 10.1145/196244.196432
[8] Md S. Uzzal, Md K. Hossen, & A. Ahmad. (July 2017).
Crosstalk Noise modeling analysis for RC Interconnect
in Deep Sub-micron VLSI Circuits. Communications on
Applied Electronics (CAE): 2394-4714. 7(4).33-38.
[9] Sapatnekar, S.,S. (June1994). R C Interconnect
Optimization under the Elmore Delay Model. Paper
presented at 31st Design Automation Conference.
IEEE. DOI:10.1145/196244.196430.
[10] Huang, X., Liu, K., Huang X., & He, Z. (Jan 2019). An
Effective Method for Interconnect Delay Optimization
of AICS. in DEStech Transactions on Engineering and
Technology Research. 255- 258. DOI:
10.12783/dtetr/ecae2018/27740.
[11] Pillage, L, T., & Rohrer, R., A. (April 1990). Asymptotic
Wave Evaluation for Timing Analysis. in IEEE
Transactions on Computer- Aided Design. 9(4). 352-
366.

Novel Global Elmore Delay Optimized Model with Improved Elmore Delay Estimation to Reduce Power Consumption and Delay

More Related Content

Similar to Novel Global Elmore Delay Optimized Model with Improved Elmore Delay Estimation to Reduce Power Consumption and Delay (20)

More from IRJET Journal (20)

Recently uploaded (20)

Novel Global Elmore Delay Optimized Model with Improved Elmore Delay Estimation to Reduce Power Consumption and Delay