You are on page 1of 14

International Journal of Psychosocial Rehabilitation, Vol.

24, Issue 10, 2020


ISSN: 1475-7192

Efficient Overburden CDMA For NoC


1
B. Mahender, 2N. Arun Vignesh, 3D. Jayanthi, 4V. AyyemPillai

Abstract-- On-chip interconnect networks causes congestion which affects the performance in
current system-on-chips. Code division multiple access (CDMA) has rigid delay, trimmed intervention
expense and superior transfer capacity. Because of this features, it’s been projected for On-Chip
crossbars. In CDMA, the standard division is facilitated in the secret language bit by appointing a
bounded quantity of unrelated(orthogonal) scattering codes of N chip range to the processing elements
distributing the link. In the existing work overburden CDMA communicate (OCI) to improve the aptitude
of CDMA based NoC (network-on-chip) crossbars by escalating the adding up to of good dispersion
codes. In this paper, the overburden CDMA implemented and the normal adder is replaced with Han
Carlson Adder (HCA) in OCI Crossbar at the time of addition. This extension of HCA gives high-speed
parallel addition, reduction in power and area. Analogized with the existing OCI on a Cyclone-2 FPGA
kit, the serial OCI attains greater transmission capacity, 30% less resource consumption, 25% power
retaining and less delay, while the parallel OCI Crossbar attains N times greater transmission capacity,
takes larger area and power. The OCI crossbar with HCA gives better results than the standard one.

Keywords--- CDMA interconnect, NoC, Overburden CDMA Crossbar, Han Carlson Adder

I INTRODUCTION

“On-chip” connections overwhelmingly contact impressive gross region, execution, along with the authority
confirmation of current facilitated circuits like ‘system on-chips’ (SoCs). Ever escalating the message on high
deteriorates the snap up reached by counterpart evaluating as claimed by Amdahl’s act [1]. Consequently, on the
increase productive eminent pursuance integrated links plays a major role in assessing technical applications. On-
chip networks are the on the whole ascendible link pattern that is apt for addressing a mixture of submission requests
and greet uncommon operation parameters of heavy workloads [2], including delay via adaptive switching [3],
throughput via increased way diversity [4], the power emission by normalizing NoC destination workloads [5], and
free nature by processing duration design [6]. In NoCs, information is treated as packets, at one time as on-chip
processing essentials (PEs) are careful as complex nodes consistent by switches and routers. On chip networks grant
an adaptable result for great on-chip systems, but they show enlarged faculty eating and significant source overheads
[7].

For enabling the concurrent channel access, here we use a different medium dispensing skill called Code-division
manifold access (CDMA) In CDMA method, every one transmit-receive (TX-RX), two of a kind, is assigned an only
one of its kind bipolar scattering secret language and facts drape from every part of transmitters are added in a

1, Department of Electronics& Communication Engineering,GRIET, Hyderabad,Telangana,India.


2Department of Electronics& Communication Engineering,GRIET, Hyderabad,Telangana,India.
3, Department of Electronics& Communication Engineering,GRIET, Hyderabad,Telangana,India
4, Department of Electronics& Communication Engineering, GRIET, Hyderabad,Telangana,India

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1138
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

preservative reporting link. In conventional CDMA spreading codes are an unrelated and the unrelated codes cross-
correlation results nil and that makes the collector in CDMA to the right to figure out the obtained added value from
a correspondent decoder. In CDMA systems medium dispensing is done by the Walsh-Hadamard unrelated codes.
CDMA advanced as an on-chip be integrated distribution procedure for in cooperation automobile and NoC
interconnect architectures [8]. Lots of return of via CDMA for on-chip interconnects rope in abridged brawn
exhaustion, fixed interaction remission, and abridged structure complication [9]. Comparing with other technologies
CDMA switch has low wiring problem and overhead. Overburden CDMA is one of the popular access method
spread out in wireless media, someplace the various user distributions the contact means is elevated by greater than
ever the add up to of usable dispersion codes with the cost of greater than ever ‘MAI’ (multiple access interference)
[10]. Theory of overburden CDMA is able to be practical to on-chip interlinks which improves the interlock ability
[11].

Adders are deep essentials in calculation operations. Dual adders are utilized as a position of a digital route for
addition, multiplication operations and for hovering aim multiplication and division. For that reason, adders are
supporting gears and civilizing their show is one of the key provocation in digital designs. Computer reckoning
algorithm has traditional drop bounds on a region and lingers ofn-chip additions, effective earlier changes
straightforward with the summation area, and its terminal possessed with an O(log2 n )nature.

This paper described in six sections, The On-chip CDMA interconnects and Han Carlson adder related work
given in II. Preliminariesregarding overburden CDMA in cellular media, conventional CDMA Crossbar switch, code
design for OCI, parallel and serial OCI designs and components described in Section III. Han Carlson Adder in
Section IV. Results and comparative analysis in Section V. Conclusion in Section VI.

II RELATED WORK

Utilizing CDMA is like an avenue contact plan in crossbar switches give propitious features like the set
undertaking delay and upward in depths negotiation. The scalable CDMA-based outermost channel to lower the
equivalent assigning position and point-to-point(PTP) links along with escape elevated TDMA controller [12]. The
indicated method decreases joi deem once second-hand at the linking of many surroundings to numerous PEs
because the numbers from the surroundings are joined and dispatched on smaller number lines.The encourage in the
undertaking delay awaited to facts expanding is customary for the reason that peripherals habitually drive at lesser
iterations over control processing elements(PE). A new CDMA channel operation and a TDMA come apart
transaction channel are analogized in [8]. A CDMA situated NoC and PTP bidirectional ring situated NoC are
compared in [13], and the difference exhibits the CDMA NoC’s preset numbers passing on latency is parallel to the
most excellent justification delay of the PTP constant duct width. The rigid figures handing over remission of the
CDMA situated NoC is attributed to concomitant be integrated allocation by the association nodes. A stratified

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1139
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

CDMA situated star NoC and the CDMA router linking done in star to star along with a star to mesh topologies
andanalyzed in terms of fair network large hierarchy topologies in [14], [15].

For a 2-D on-chip network this CDMA situated multicast exchange is employed in [16]. Because of code space
multiplexing, The CDMA situated exchange agrees parallel data communication. The multicasting hop count
decreased by this method and it agrees the target PEs made by packets concurrently, that is favored for practical
uses. A CDMA of 14 node set-up has shown in [17]. The Tx-Rx duo is set by the spreading codes is effective depend
on the call from every node. Two building blocks are proposed in this CDMA situated network, they are namely
serial and parallel CDMA networks. The CDMA takes every data in spreading code is given to one-time period in
the serial network where all the data is given that same clock period in the parallel network. The CDMA situated
parallel and serial networks possess been analogized with a classical CDMA, TDMA bus and a net placed NoC. The
uniform CDMA’s bandwidth is greater to net placed NoC’s throughput for the consistent network area due to
concurrent behavior of CDMA [11].

There is quantity of adder architectures are available, the counterpart affix architectures of adder circuits bear
meeting place stand straight to the drive too fast by that effective yield is gained. The time O(log n) gives the n-bit
information total [18]. The foundation for parallelism with move give impression of being prematurely accumulation
technique done by two people namely Weinberger along with Smith. The base of parallel prefix adders that the
calculation done in prefixes are regarded towards regularity and it performed in parallel, this is established by P.M.
Kogge along with H.S Stone.

III PRELIMINARIES

This section discusses, in wireless media, the role of this overburden CDMA, the conventional CDMA Crossbar
switch are presented [12]. The code structure for OCI, building blocks of OCI in serial along with parallel manner
are described.

III.I. Overburden CDMA role in wireless transmissions


DSSS-CDMA is the foremost loom as average allocation of radio communications someplace a settle of opposite
scattering codes poised of an issue of chips in segment N are accumulated by the sent message bits just as all
information morsel is unfolded in N time periods. An only one of its kind scattering convention is given to each Tx-
Rx two of a kind division the transfer channel. Information flows of user’s allocation the bus are multiple and
concurrently sent to the addition medium link. De-spreading is obtained by giving the correlation business to the
gained sum, anywhere all accepter knows how to obtain its facts by doing correlation with the allotted diffusion
code. Orthogonal nature among the distribution codes bonds exceptional finding of each programming expected in
the avenue addition by utilizing the associative and distributive conventions of the tallying sum procedure accepted
by the transmission bus. In radio transmissions, true locations of the standard addition are affected due to

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1140
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

indiscriminate properties increases, namely multipath, fading, noise and which also causes for getting the more error
in bits(BER) of received information. In an unfavorable, in the scattering convention array, amount of unrelated
codes is customarily narrow to the scattering secret language time taken by N, which decrease the network
exploitation ability. In wireless media novel, introduction of Overburden CDMA to expansion in the quantity of
scattering codes aside summing non-orthogonal(related) codes which perhaps spotted at accepter part[10]. Escalating
the bus consumption advance at overhead of lighten the unrelated nature necessities of the unroll codes and
escalating MAI and this gives hike in BER.

In wireless media, this projected overburden CDMA’s spreading bits are guided with intricate acceptor designs
wear and tear of multiuser detection as a replacement for of the easy matched filter or correlator accepter employed
in plain direct sequence spreadspectrum code division multiple access. Here, the CDMA situated NoC ability
increased by applying this overburden CDMA approachin wireless media [11].

III.II. Classical CDMA Crossbar Switch


The router’s material layer which is based on the conventional CDMA conferred in [12]. The switch is made up
of a quantity of XOR encoders, a physical link addition circuitry and decoders which are situated on the
accumulator. For encoding, Walsh scattering codeset produces the binary symmetric code of length N, is encoded
with the conveyed message data bit and forwarded out in serial manner. It tells that a 1-bit enlargement done in a
span of N time periods. Accordingly, the frequency ft of the crossbar affairs and managing clock frequency fc , both
denoted as ft = fc/N

For Walsh scattering codes, the number of transmitter(TX) and receiver(RX) ports taking CDMA switch equal to
the M = N–1. Crossbar sharing done in serially by the all transmitted PEs and they are summed, the sum sent to the
decoder part of the circuit filling the Rx ports. The binary signaling and encoding has great execution, responsibility,
and its intrinsic inherent carry by digital policies, this is reason for using it to implement the bus addition over the
multilevel signaling. The message bits through the CDMA crossbar exchange is in terms of equation:

(1)

Here, m bit digital value is S(i) produces bus total in the duration of ith time period. the width of crossbar m
=[log2 M], message bits d(j) is from the jth encoder, ⊕ gives XOR behavior and C0(j,i) gives ithpart of jth
symmetric scattering code. The addition circuitry in conventional CDMA crossbar with inputs M = N−1 bits along
with m = log2N = [log2 M] yield information.

The correlator decoder understands by the two accumulators. As demonstrated by the dispensed CDMA code,
zero accumulator gets obtained total when the existing chip denomination is ‘0’ and one accumulator gets the sum
when the denomination is ‘1’. This is practically identical towards magnifying crossbar total with ±1.The CDMA
sum along with crossbar total for the different spreading code logical conditions are shown in [11]. Differentiating
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1141
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

the two accumulators particularly shows the encoded data; if the data of zero accumulator is greater when its
compared with 1 accumulator, then the issued message bit is ‘1’, otherwise bit is ‘0’. Hence, the association action
able to choose the encoded message with null faults in view of dismissing subjective effects. The magnification
situated correlator is replaced with a inclusion situated one, this is the prime choice of accumulator decoder.

III.III. Overburden CDMA interconnect


Because of focus overloading the main difference among the overloaded and conventional CMDA network
device is M>N-1. Network interfaces are attached by each PE which transmits and obtain NI Modules. While packet
sending from each PE, to store in transmit NI Queue every single packet splits into flow control digits. The arbiter
chooses M gaining flow control digits from leading NI Queue and sent at the time present activity. To avoid
disputes, elected flow control digits need to have dedicated target address, as stated by router's priority plan gainer
from two disputing flow control digits is elected. Employed prime plan is the stable gainer that hold of all prime
plans; only one sender and receiver will be provided with a spreading code, is accepted as true to begin the encoding.
As soon as it is done, the network device(router) allocates CDMA codes to every NI transmit and NI receive. NI's
having void queue's or dispute destinationsare allocated with all zero CMDA codes so they will not give a
contribution to MAI to CDMA channel total. At a later, in encoder part, Flow control digits from every single NI are
divided by CDMA Codes. N chips of Information is divided, and N is length of CDMA that is same as amount of
time period mono crossbar activity. All decoders will receive a sum sequentially which is the addition of smear
information chips from every encoder by CDMA cross bar adder. Using counter, decoding or encoding procedure
continues till “N” frequencies occurs at same time [11].

III.III.I. Top-level design of OCI Crossbar


Primary goal of the active manuscript is growing quantity of storage elements by distributing typical CDMA
cross bar, throughout the time preserving the density of the theory consistently by means of uncomplicated encoding
design and depending on the decoder which is situated on accumulator having nominal differences. To reach this
objective, a few alterations to the typical CDMA cross bar are ahead in position. For single bit interdependence,
Figure.1 characterizes the esteemed design of OCI cross bar. The unchanged architecture is imitated for a multi-bit
CDMA switch. The CDMA switch split by transmitter-receiver ports of length M, that uses computational twofold
addition circuitry having M bit information sources and yield is m-bit, which shows m =log2M for adding spreading
message from the Tx ports. The mentioned and duct designs take the advantage of adder. For facilitating the message
spreading and dispreading, encoder and decoder envelope coupled by all PEs. The cryptogram task along with
intervention undertakings take usage of controller stall. The XOR operation is used for unrelated dissipating codes
and an AND operation is used for related dissemination codes. The operation of AND as pursues, if the Rx message
bit is ‘0’, it assigns the bunch of zeros for the complete dissemination period, here MAI(multiple access interface) on
bus will not occur; if the Rx message bit is ‘1’, then the encoder assigns an related dispersal code. Along these lines,
being the encoder is AND, the developed MAI dissemination cod e gives its value ‘1’or ‘0’. The regular CDMA
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1142
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

crossbar’s XOR encoder is not used for OCI encoding due to supplements of dispersal codes, thus the MAI caused
by XOR gates for both one and zero. For both unrelated and related widening, an amalgam encoder introduced and
its with XOR, AND gates, along with mux as depicted in Fig.1. The execution of related along with related
information by dual decoder models.

III.III.II. Code structure for OCI


The family of Walsh–Hadamard dissipating codes highlighted peculiarity which empowers CDMA to be
coordinated over-burdening. The contrast between any continuous course wholes of information advance with
related widening bits to the different quantity of transmitter and receiver set is constant, i.e. M, careless dispersal
information. These home incomes that for the N−1 TX-RX sets utilizing affective Walsh unrelated bits, particular
may encode extra N−1 fact codes in back to back contrasts among the N lengthchips making unrelated code. By the
use of this rule facilitate 100% adding of related dispersal codes, that makes efficient classical CDMA crossbar. In
this way, in related encoder, message for transmission is ‘1’ an all alone appropriation piece at an individual vacancy
in the spreading stage is extra to the lead aggregate, which causes the continuous expansion change to go amiss. The
related codes pretend to be the TDMA flagging arrangement as both cryptogram is made out of a particular sign of
"1"sent in an express schedule opening. The encoding or decoding plot offered in this address give a novel think
about that empowers concurrence among code and time division multiple access motions for the steady joint
channel. In this way, the private encoder is known as TDMA overburden on OCI (T-OCI). An encoding/deciphering
case of dual TDMA Overburden CDMA Interlink bits to diffusing arrangement guidelines length of N=8 is
demonstrated in Fig.1. A different quantity of unrelated codes ought to be utilized all the while to diversion save the
dimension difference material merchandise of Walsh codes. The scientific establishments for the entirety encoded
conspire and decoded plot in Ahmed el at [1].

III.III.III. Building Blocks of OCI crossbar


For every crossbar and associated duct designs, there are duo alternatives are acknowledged. The actualization of
duct designs is to improve the crossbar actualization recurrence, and the data transfer capacity by collection
impractical duct registers which are used for diminishing the crossbar crucial range. The description of architectures
of OCI crossbar are explained below.

Crossbar Scrutinizer: the initial for every crossbar exchange, this scrutinizer appoints scattering bits towards
various encoders. There is no change among the crossbar exchange in terms of, because the unrelated codes are
settled at acceptor side which are appointed at the transmission side. Consequently, the switch port begins the
connection towards acceptor side it directs, here encoder should be appointed for scattering code which also must
equals the acceptor side decoder. A particular decoder gets the appeal from various dual ports, scrutinizer permits
single entryway and rejects another as indicated by the predefined intervention conspire [13]. The crossbar
scrutinizer circulates concord signs towards Tx-Rx storage places with coordinating scattering bits to allow Tx
encoders along with Rx decoders.
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1143
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

2) Hybrid Encoder: The encoder is crossover, it be equipped for encoding both the information to be specific
unrelated and non-orthogonal. The XOR and AND operation done between Rx messages and dissipating bits for
creating unrelated/related mushroom information, individually. On the basis of bit model appointed to an encoder, a
multiplexer picks between the unrelated and related contributions as shown in Fig.1(a). The encoder is copied ‘N’
time for the P-OCI crossbar.

3) Crossbar Adder: The quantity of crossbar transmitter and receiver ports coordinated towards M=2(N-1) for
dispersing codes of range N. On the basis of TDMA-OCI crossbar encoding method, when we apply ‘1’ to the adder,
its mutually exclusive among the related ports.There are assured zeros of N-2, as surrounded by 2(N-1) information
to the addition circuitry, where maximal quantity N of ‘1’ chips. Accordingly, one Mux is introduced for the election
of one admission from the related TDMA encoded information bits and removes left out bits as ‘o’. Hence, the
admissions which addition circuitry includes N-1 from unrelated encoders along with ‘one’ from mux, so it’s clear
that its having N admissions as portrayed in Fig.1(d). The quantity of considered essential phases of registry to duct
the addition circuit is log2N, this shown in Fig.1(d). For parallel OCI, these crossbar addition circuitries are repeated
for N times.

(d)

(e)

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1144
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

(a) (b) (f)

(c)

F ig 1. Top level design of OCI Crossbar. (a) TDMA-OCI/Parallel-OCI hybrid encoder. (b) TDMA-OCI non-
orthogonal decoder. (c) Parallel-OCI non-orthogonal decoder. (d) TDMA-OCI pipelined crossbar tree adder
(duplicated N chances for P-OCI crossbar). (e) Parallel-OCI orthogonal decoder. (f) TDMA-OCI orthogonal
decoder.

Fig .2 Encoding /Decoding of orthogonal codes and two non-orthogonal code.


4)Custom Decoder: For various CDMA decoding methods, the decoder models are 4 in number. They are
namely 1). Unrelated TDMA OCI, 2). Unrelated Parallel OCI, 3). Overburden(related) TDMA OCI, 4). Overburden
Parallel OCI. An accumulator’s execution of correlation acceptor is expressed in terms of unrelated TDMA-OCI
decoder. For unrelated message non-propagation, N-1 accumulator decoders are called in the whole CDMA crossbar
models. The up-down accumulator replaces the dual various accumulators and the outcome is divergence of dual
accumulators of ordinary CDMA decoder as depicted in Fig.1(f).

As reporting to the non-propagation code chip, the crossbar total parameter’s addition/subtraction done by the
accumulator and restarts. The decoding bit is depending on the sign bit of the accumulator’s output, that +ve clue
gives ‘1’ and –ve clue gives ‘0’ decoding bits. An accumulator presented as parallel addition circuitry due the
Parallel-OCI unrelated decoder contrasts from contrasts against TDMA-OCI unrelated decoder in getting addition
total parameter simultaneously not progressively as portrayed in Fig.1(e). A two bit register is used in TDMA-OCI
overburden(related) decoder to accumulate the LSB (least significant bit) s of 2 added parameters as shown in
Fig.1(b), at S(0) is primary one and S(j-N+1) is 2nd, that the j represents the quantity of TDMA-OCI decoders
(N≤j≤2N−2) and widening code length N. This 2-bit information urged XOR gate, that unwinds non-
orthogonal(related) accomplish message. Fig.1(c) shows that the TDMA OCI decoder is copied N chances for
executing Parallel OCI. There is no need of two-bit register in P-OCI related decoder due to all the bit values remain
in a time. The N-1 unrelated and related decoders exist in TDMA OCI and Parallel OCI crossbar designs.
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1145
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

IV HAN CARLSON ADDER

In Overloaded-CDMA crossbar we have an adder circuitry to add the encoded messages together to produce a
sum signal. We are performing additions with normal adder which takes the large delay for parallel additions. To this
extent,we are replacing normal adders with the design of Han Carlson adderto achieve high-speed parallel addition
and it also reduces the area and power. In prefix accumulation we wastage three stages to enlist the total in pre-
processing, prefix-processing and post-preparing.

The aim of combining is to reckon the sum S, for the two parts A and B of length n in binary. The first phase, for
n bit sum adder calculates the generate(g) and propagate(p) phases per bit of the operands according to the following
g i  ai .bi
equations. (2)
pi  ai  bi

From the equation 2, prefix addition performed. The prefix operation ʘ,is described as :

(gi, pi) ʘ (gj ,pj) = (gi,+ pi • gj, pi •∙ pj) (3)

For every prefix activity, the operation associates the 2 AND operations along with 1 OR operation. Here two
vital characteristics namely idempotency and associativity shown by the prefix operation. Therefore, it gives the
following:

(gm…i ,pm….i, ) ʘ ( gi…n ,pi….n) = ( gm…n, pm…...n) (4) (gm…i ,pm….i, ) ʘ( gi…n ,pi...n)=( gm..j, pm…j ) ʘ(
gi…n, pj…n )(5)

The calculation of carry for any bit point is given as below equation (5), as a loop of prefix operations:

(gi…o ,pi….o) = (gi-1,pi-1) ʘ (gi-2,pi-2)………(g1,p1)ʘ (go,po) (6)

Equation (7) tell that carry generated from equation (6),

ci = gi….o+pi….o•co(7)

The last phase, the carry bits and propagate bits are used to calculate the sum. This shown in equation as:

si  pi  ci
(8)

From equation (8), it’s clear that the post addition operation performed with the propagated bits and carry bits.
The prefix operators are known as ‘parallel prefix adder’ which are required to build the adders. For the calculation
of group g’s (gi…j) and group p’s (pi…j), they apply the associativity and idempotency property of the prefix
operators.

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1146
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

Fig .3 Han Carlson Adder


A Kogge-Stone adder consequences in an allocation of side-by-side networks and for this reason bottleneck,
unchanging still the architecture regulates the fan out to unification and calculates the carry in O(log n) time [19]. On
the other side, prefix per node in graph is calculated by a Brent Kung adder, restricting the reasonable fan-out to
integrity, but without the using of wires. The Han Carlson figure of a 16-bit analogy prefix adder uses a particular
Brent Kung platform Kogge-Stone plan and terminated by a further Brent-Kung show for the terminal playhouse of
the prefix calculation.

Further processing after the calculation of the parallel prefixes is like for completely the similar prefix adders.
The prefix calculation phase of a 16- bit Han Carlson adder is exposed in Fig. 3. It t is experiential from Fig. 3 there
are five prefix calculation phases for the Han Carlson adder, which is one additional than the Kogge-Stone design
(log216=4) for the unaffected word size.

Fig.4 HCA implementation in Tree adder


The Fig.4 shows that HCA in OCI crossbar, the normal carry select adder is replaced with HCA. This gives
better results than the previous addition circuitry.

V RESULTS AND COMPARATIVE ANALYSIS

This section discusses, the performance evaluation outcomes of the OCI Crossbar with normal adder and the
proposed Han Carlson Adder are presented.

1. OCI Crossbar evaluation:


In this, normal OCI CDMA and OCI with HCA are compared and these crossbars calculated for widening bits of
ranges N {8,16}. Crossbar designs with the different quantity of ports, whole fulfilment measurements are
assimilated to M quantity crossbars. The estimation outcomes, covering capital usage conveyed as quantity of LUTs
(Look Up Tables) along with flip flops(FF) for each port per port, ultimate crossbar recurrence, effective power
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1147
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

exhaustion for each port and transmission limit are shown in Fig. 5. The Fig.5(a) shows that, OCI crossbar with
HCA’s resource usage is 30% less than the ordinary OCI crossbar for a N length widening bits. As N spreading code
length raises, resource usage for each port is increased, because of gaining crossbar intricacy.

The working recurrence reduction done with crossbar adder’s crucial way range for all mentioned designs. A
confined length of N widening codes of different CDMA crossbar, unrelated propagating and non-propagating
designs are same and related information encoders and decoders are functioning equal to unrelated widening circuit
of small crucial way range. The admission part of addition hardware equivalent to the quantity RX ports M, that
changes by CDMA crossbar model. The crossbar recurrence of OCI with HCA crossbar is minimal higher when it
compared with OCI with normal addition circuitry because of improving the addition design for a constant widening
bit length ‘N’ shown in Fig.5(b). The Fig.5(c) shows that, gaining crossbar data transfer capacity in OCI with HCA
than that of OCI with normal adder, it satisfies the recurrence reduction with expanding N from overburdening. The
chart balanced transmission capacity is drawn to a one for each port linked through CDMA crossbar. The
upgradation of CDMA crossbar with HCA data transfer capacity over the traditional CDMA crossbar for Parallel
OCI along with TDMA OCI crossbar is striking for constant N. By and large, the CDMA crossbar data transfer
capacity, i.e BW is expressed by the progressive condition:

M
BW  Wf c (9)

From condition (9) bandwidth, that thickness of port is W in bits, the crossbar time recurrence is fc , M is quantity
of crossbar ports, quantity loop for encoding the one bit information from whole ports is  [11].

The effective power diversion for each port assessed with a tool called Xilinx Vivado to one crossbar activity,
25% is diminished for OCI with HCA against the TDMA-OCI crossbar because of the offered limit upgrade for a
widening code N of length as shown in Fig.5(d). In any case, because of the expanded zone and parallel encoding
disentangling of Parallel OCI crossbar, the effective power diversion is less when it compared with OCI with the
normal adder. The power diversion for each port raises for whole CDMA crossbars by expanding N, because of
extended magnitude and intricacy of crossbar elements.

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1148
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

Fig.5 Execution outcomes of OCI crossbar for N range spreading code. (a) resource utilization(b) Frequency
(c)Bandwidth (d)Dynamic power in mWper port.
2. OCI for Network On Chip: Experimental Evaluation

For complete operation of Network On Chip takes the importance of OCI crossbars, the 5 OCI switches are used
to develop the 65-hub star topology, in 13 PEs every single is associated with OCI switch of N=8, a Spatial division
multiple access’ basic switch is linked by 5 OCI switches [11]. The effectiveness of OCI with normal adder and with
HCA on Cyclone-2 FPGA kit as shown in table.1.

Table. I Experimental results:

The table I provides the resource utilization expressed in the several Look Up Tables (LUTs), flip flops (FF) and
input-output blocks, greatest latency and dynamic power consumption. Comparing the numerical values, we can see
that area is reduced by 30%, 12% less delay and 25% reduction in power consumption. The throughput  which
can find out as:

Nc  Nb  N p (10)

tc

Where, quantity of synthesis clock periods is Nc,Nbis the quantity of bits/packet, the quantity of packets obtained
by destination PEs is Np, the clock period is tc. The throughput of the OCI with HCA is higher than OCI with normal
adder because of its lower time period.

VI CONCLUSION

The CDMA concept has been projected for on-chip communication and the limited number of N-chip unrelated
widening codes shares the interconnect.In this work, we advance with overburden CDMA, NoC router’s physical
layer is enabled by this overburden CDMA crossbar. In overburden CDMA, the transfer drain is overburdened with
related codes to become more intense the channel ability. In this work, the normal adder replaced with Han Carlson
Adder(HCA) in OCI Crossbar. This thought of HCA because of its parallel addition concept and less complexity.
The occurrence of the OCI crossbars is analyzed with and without HCA in crossbar at the time of addition and
examined on a Xilinx Cyclone-2 FPGA kit. The effective power by using Vivado tool is 25% diminished for the OCI
crossbar, where HCA is used for crossbar addition. The OCI with HCA crossbar drains 30% less resources than that

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1149
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

of ordinary OCI system and the delay is reduced by 2 nsec. Observing the heaviness of CDMA links and CDMA’s
improvement methods intended as part of the preceding impending study points.

REFERENCES

[1] Manoharan, Rajesh, et al. "Selection of Intermediate Routes for Secure Data Communication Systems using
Graph Theory Application and Grey Wolf Optimization Algorithm in MANETs." IET Networks (2020).

[2] Rajesh, M., Gnanasekar, J.M. Path Observation Based Physical Routing Protocol for Wireless Ad Hoc
Networks. Wireless Pers Commun 97, 1267–1289 (2017). https://doi.org/10.1007/s11277-017-4565-9

[3] Rajesh, M. Streamlining Radio Network Organizing Enlargement Towards Microcellular


Frameworks. Wireless Pers Commun (2020). https://doi.org/10.1007/s11277-020-07336-9

[4] K. Asanovic et al., “The landscape of parallel computing research: A view from berkeley,” Dept. EECS, Univ.
California, Berkeley, CA, USA,Tech. Rep. UCB/EECS-2006-183, 2006.

[5] P. Bogdan, “Mathematical modeling and control of multifractal workloads for data-center-on-a-chip
optimization,” in Proc. 9th Int. Symp. Netw.-Chip, New York, NY, USA, 2015, pp. 21:1–21:8.

[6] Z. Qian, P. Bogdan, G. Wei, C.-Y. Tsui, and R. Marculescu, “A trafficaware adaptive routing algorithm on a
highly reconfigurable network-onchip architecture,” in Proc. 8th IEEE/ACM/IFIP Int. Conf. Hardw./Softw.
Codesign, Syst. Synth., New York, NY, USA, Oct. 2012, pp. 161–170.

[7] Y. Xue and P. Bogdan, “User cooperation network coding approach for NoC performance improvement,” in
Proc. 9th Int. Symp. Netw.-Chip, New York, NY, USA, Sep. 2015, pp. 17:1–17:8.

[8] T. Majumder, X. Li, P. Bogdan, and P. Pande, “NoC-enabled multicore architectures for stochastic analysis of
biomolecular reactions,” in Proc. Design, Autom. Test Eur. Conf. Exhibit. (DATE), San Jose, CA, USA,
Mar. 2015, pp. 1102–1107.

[9] S. J. Hollis, C. Jackson, P. Bogdan, and R. Marculescu, “Exploiting emergence in on-chip interconnects,”
IEEE Trans. Comput., vol. 63, no. 3, pp. 570–582, Mar. 2014.

[10] S. Kumar et al., “A network on chip architecture and design methodology,” in Proc. IEEE Comput. Soc.
Annu. Symp. (VLSI), Apr. 2002, pp. 105–112.

[11] R. H. Bell, C. Y. Kang, L. John, and E. E. Swartzlander, “CDMA as a multiprocessor interconnect strategy,” in
Proc. Conf. Rec. 35th Asilomar Conf. Signals, Syst. Comput., vol. 2. Nov. 2001, pp. 1246–1250.

[12] B. C. C. Lai, P. Schaumont, and I. Verbauwhede, “CT-bus: A heterogeneous CDMA/TDMA bus for future
SOC,” in Proc. Conf. Rec. 35th Asilomar Conf. Signals, Syst. Comput., vol. 2. Nov. 2004, pp. 1868–1872.

[13] S. A. Hosseini, O. Javidbakht, P. Pad, and F. Marvasti, “A review on synchronous CDMA systems: Optimum
overloaded codes, channel capacity, and power control,” EURASIP J. Wireless Commun. Netw., vol. 1, pp.
1–22, Dec. 2011.

[14] Khaled E.Ahmed and M.Farag, “Overloaded CDMA crossbar for network-on-chip” IEEE trans. Very Large
Scale Intgr.(VLSI) syst,vol.25,No 6, june2017.

[15] T. Nikolic, G. Djordjevic, and M. Stojcev, “Simultaneous data transfers over peripheral bus using CDMA
technique,” in Proc. 26th Int. Conf. Microelectron. (MIEL), May 2008, pp. 437–440.

[16] X. Wang, T. Ahonen, and J. Nurmi, “Applying CDMA technique to network-on-chip,” IEEE Trans. Very
Large Scale Integr. (VLSI) Syst., vol. 15, no. 1o, pp. 1091–1100, Oct. 2007.
Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1150
International Journal of Psychosocial Rehabilitation, Vol. 24, Issue 10, 2020
ISSN: 1475-7192

[17] D. Kim, M. Kim, and G. E. Sobelman, “CDMA-based network-on-chip architecture,” in Proc. IEEE Asia–
Pacific Conf. Circuits Syst., vol. 1. Dec. 2004, pp. 137–140.

[18] D. Kim, M. Kim, and G. E. Sobelman, “Design of a high-performance scalable CDMA router for on-chip
switched networks,” in Proc. Int. SoC Des. Conf, Nov. 2005, pp. 32–35W. Lee and G. E. Sobelman, “Mesh-
star hybrid NoC architecture with CDMA switch,” in Proc. IEEE Int. Symp. Circuits Syst. (ISCAS), May
2009, pp. 1349–1352.

[19] K. E. Ahmed and M. M. Farag, “Overloaded CDMA bus topology for MPSoC interconnect,” in Proc. Int.
Conf. ReConFigurableComput. FPGAs (ReConFig), Dec. 2014, pp. 1–7.

[20] P. M. Kogge and H. S. Stone, “A parallel algorithm for the efficient solution of a general class of recurrence
equations,”IEEE Trans. Computer., vol. C-22, no. 8, pp. 786–793, Aug. 1973.

[21] T. Han and D. Carlson, .Fast area-efficient VLSI Adders, In Proc. 8th Symp.Comp. Arithmetic, Sept. 1987, pp.
49.56.

Received: 27 Feb 2019 | Revised: 20 Mar 2019 | Accepted: 30 Apr 2020 1151

You might also like