International Journal of Engineering and Advanced Technology (IJEAT)
ISSN: 2249 9!" #olu$e%&" I''ue%!" June 2(&2
I$)le$entation of an Efficient *ulti)lier +a'ed on #edic *athe$atic' ,'ing E-A Tool .u'h)alata #er$a" /0 /0 *ehta
Abstract:% A high speed processor depends greatly on the multiplier as it is one of the key hardware blocks in most digital signal processing systems as well as in general processors. This paper presents a high speed 8x8 bit Vedic multiplier architecture which is quite different from the Conventional method of multiplication like add and shift. The most significant aspect of the proposed method is that the developed multiplier architecture is based on Vertical and Crosswise structure of Ancient !ndian Vedic "athematics. !t generates all partial products and their sum in one step. This also gives chances for modular design where smaller block can be used to design the bigger one. #o the design complexity gets reduced for inputs of larger no of bits and modularity gets increased. The proposed Vedic multiplier is coded in V$%& 'Very $igh #peed !ntegrated Circuits $ardware %escription &anguage( synthesi)ed and simulated using *%A '*lectronic %esign Automation( tool + ,ilinx!#*-..-i. /inally the results are compared with Conventional multipliers to show the significant improvement in its efficiency in terms of path delay 'speed(. The high speed processor requires high speed multipliers and the Vedic "ultiplication technique is very much suitable for this purpose. !ndex Terms:% Architecture 0ipple Carry '0C( Adder "ultiplication Vedic "athematics Vedic "ultiplier 'V"( 1rdhava Tiryakbhyam #utra I0 INT12-,3TI2N Multipliers are extensively used in Microprocessors, DSP and Communication applications. For higher order multiplications, a huge number of adders are to be used to perform the partial product addition. The need of high speed multiplier is increasing as the need of high speed processors are increasing. igher throughput arithmetic operations are important to achieve the desired performance in many real time signal and image processing applications. !ne of the "ey arithmetic operations in such applications is multiplication and the development of fast multiplier circuit has been a sub#ect of interest over decades. $educing the time delay and po%er consumption are very essential re&uirements for many applications0 'n the past multiplication %as implemented generally %ith a se&uence of addition, subtraction and shift operations. T%o most common multiplication algorithms follo%ed in the digital hard%are are array multiplication algorithm and (ooth multiplication algorithm. Due to the importance of digital multipliers in DSP, it has al%ays been an active area of research. )edic mathematics is the name given to the ancient system of mathematics, %hich %as rediscovered from the *anu'cri)t received on June 22" 2(&2. *r'0 .u'h)alata #er$a, M.*. Scholar +CT,-, SSC*T, (hilai, Chhattisgarh, 'ndia. -r0 /0 /0 *ehta, .or"ing as Professor / ead of Department in Computer Science / *ngineering at SSC*T, (hilai, Chhattisgarh, 'ndia. ancient 'ndian scriptures bet%een 0100 and 0102 by 3agadguru S%ami Sri (harati 4risna Tirtha#i +022560178-, a scholar of Sans"rit, mathematics, history and philosophy. The %hole of )edic mathematics is based on 07 )edic sutras, %hich are actually %ord formulae describing natural %ays of solving a %hole range of mathematical problems 90:. The paper is organi;ed as follo%s. Section '' describes the basic methodology of )edic multiplication techni&ue. Section ''' describes the proposed multiplier architecture based on )edic multiplication and the generali;ed algorithm for <x< bit )edic multiplier. Section ') describes the design and implementation of )edic multiplier module in =ilinx'S*0>.0. Section ) comprises of $esult and Discussion in %hich device utili;ation summary and computational path delay obtained for the proposed )edic multiplier +after synthesis- is discussed. Finally Section )' comprises of Conclusion. II0 #E-I3 *,4TI.4I3ATI2N TE35NI6,E The use of )edic mathematics lies in the fact that it reduces the typical calculations in conventional mathematics to very simple one. This is so because the )edic formulae are claimed to be based on the natural principles on %hich the human mind %or"s. )edic Mathematics is a methodology of arithmetic rules that allo% more efficient speed implementation. 't also provides some effective algorithms %hich can be applied to various branches of engineering such as computing. A. Urdhva Tiryakbhyam Sutra The proposed )edic multiplier is based on the ?@rdhv TiryagbhyamA+algorithm-sutra.TheseSutras have been traditionally used for the multiplication of t%o numbers in the decimal number system. 'n this %or", %e apply the same ideas to the binary number system to ma"e the proposed algorithm compatible %ith the digital hard%are. 't is a general multiplication formula applicable to all cases of multiplicationVertically.'tandliterallyCrosswiseA. 't is based on a novel concept through %hich the generation of all partial products can be done %ith the concurrent addition of these partial products. The algorithm can be generali;ed for n x n bit number. Since the partial products and their sums are calculated in parallel, the multiplier is independent of the cloc" fre&uency of the processor. Due to its regular structure, it can be easily layout in microprocessors and designers can easily circumvent these problems to avoid catastrophic device failures. The processing po%er of multiplier can easily be increased by increasing the input and output data bus %idths since it has a &uite a regular structure. Due to its regular structure, it can be easily layout in a silicon chip. The Multiplier based on this sutra has the advantage that as the number of bits increases, gate delay and area increases very slo%ly as compared to other conventional multipliers. 7 ! I$)le$entation of an Efficient *ulti)lier +a'ed on #edic *athe$atic' u'ing E-A Tool B. Multiplication of two decimal numbers ! " #$% To illustrate this scheme, let us consider the multiplication of t%o decimal numbers >B> x 257 by @rdhva6Tirya"bhyam method as sho%n in Fig. 0. The digits on the both sides of the line are multiplied and added %ith the carry from the previous step. This generates one of the bits of the result and a carry. This carry is added in the next step and hence the process goes on. 'f more than one line are there in one step, all the results are added to the previous carry. 'n each step, least significant bit acts as the result bit and all other bits act as carry for the next step. 'nitially the carry is ta"en to be ;ero. 8ig0 & *ulti)lication of t9o deci$al nu$+er' 2!2 : 4; III0 T5E .12.2SE- *,4TI.4IE1 A135ITE3T,1E The hard%are architecture of >=>, 5x5 and 2x2 bit )edic multiplier module are displayed in the belo% sections. ere, ?@rdhva6 TiryagbhyamA+Verticallyand Crosswise- sutra is used to propose such architecture for the multiplication of t%o binary numbers. The beauty of )edic multiplier is that here partial product generation and additions are done concurrently. ence, it is %ell adapted to parallel processing. The feature ma"es it more attractive for binary multiplications. This in turn reduces delay, %hich is the primary motivation behind this %or". A. Vedic Multiplier for " bit Module The method is explained belo% for t%o, > bit numbers A and B %here A C a0a8 and B C b0b8 as sho%n in Fig. >. Firstly, the least significant bits are multiplied %hich gives the least significant bit of the final product +vertical-. Then, the DS( of the multiplicand is multiplied %ith the next higher bit of the multiplier and added %ith, the product of DS( of multiplier and next higher bit of the multiplicand +cross%ise-. The sum gives second bit of the final product and the carry is added %ith the partial product obtained by multiplying the most significant bits to give the sum and carry. The sum is the third corresponding bit and carry becomes the fourth bit of the final product. s8 C a8b8E The final result %ill be c>s>s0s8. This multiplication method is applicable for all the cases. 8ig0 2 The #edic *ulti)licat ion *ethod for t9o 2% +it <inary Nu$+er' The >=> )edic multiplier module is implemented using four input ,<D gates / t%o half6adders %hich is displayed in its bloc" diagram in Fig. F. 't is found that the hard%are architecture of >x> bit )edic multiplier is same as the hard%are architecture of >x> bit conventional ,rray Multiplier 9>:. ence it is concluded that multiplication of > bit binary numbers by )edic method does not made significant effect in impro efficiency. )ery precisely %e can state that the total delay is only >6half adder delays, after final bit products are generated, %hich is very similar to ,rray multiplier. So %e s%itch over to the implementation of 5x5 bit )edic multiplier %hich uses the >x> bit multiplier as a basic building bloc". The same method can be extended for input bits 5 / 2. (ut for higher no. of bits in input, little modification is re&uired. 8ig0 = <loc> -iagra$ of 2:2 +it #edic *ulti)lier B. Vedic Multiplier for $"$ bit Module The 5x5 bit )edic multiplier module is implemente d using four >x> bit )edic multiplier modules as discussed in Fig. F. DetGs analy;e 5x5, multiplicat,,,and(C F ( F ( > ( 0 ( 8 . The output line for the multiplication result is HS I S S S S S S S DetGs divide , and, ( i 7 B 5 F > 0 8 . F , > / , 0 , 8 for , and ( F ( > / ( 0 ( 8 for (. @sing the fundamental of )edic multiplication, ta"ing t%o bit at a time and using > bit multiplier bloc", %e can have the follo%ing structure for multiplication as sho%n in Fig. 5. c0s0 C a0b8 J a8b0E +>- cs> C c0 J a&b&E +F- 7; 8ig0 4 Sa$)le .re'entation for 4:4 +it #edic *ulti)lication *ach bloc" as sho%n above is >x> bit )edic multiplier. First >x> bit multiplier inputs are , 0 , 8 and ( 0 ( 8 . The last bloc" is >x> bit multiplier %ith inputs , F , > and ( F ( > . The middle one sho%s t%o >x> bit multiplier %ith inputs , F , > / ( 0 ( 8 and , 0 , 8 / ( F ( > . So the final result of multiplication, %hich is of 2 bit, S I S 7 S B S 5 S F S > S 0 S 8 . To understand the concept, the (loc" diagram of 5x5 bit )edic multiplier is sho%n in Fig. B. To get final product +S I S 7 S B S 5 S F S > S 0 S 8 -, four >x> bit )edic multiplier +Fig. F- and three 56bit $ipple6Carry +$C- ,dders are re&uired. The proposed )edic multiplier can be used to reduce delay. *arly literature spea"s about )edic multipliers based on array multiplier structures. !n the other hand, %e proposed a ne% architecture, %hich is efficient in terms of speed. The arrangements of $C ,dders sho%n in Fig. B, helps us to reduce delay. 'nterestingly, 2x2 )edic multiplier modules are implemented easily by using four 5x5 multiplier modules. 8ig0 ; <loc> -iagra$ of : +it #edic *ulti)lier '. (enerali)ed Al*orithm for + " + bit Vedic Multiplier .e can generali;e the method as discussed in the previous sections for any number of bits in input. Det, the multiplication of t%o <6bit binary numbers +%here < C 0, >, FK<, must be in the form of > < - A and B %here , C , < ....., F , > , 0 and ( C ( < K.( F> ( 0 . The final multiplication result %ill be of +< J <- bits as S C S +< J <- ....S F S > S 0 . Ste) &: Divide the multiplicand , and multiplier ( into t%o e&ual parts, each consisting of 9< to +<L>-J0: bits and 9<L> to 0: bits respectively, %here first part indicates the MS( and 8ig0 ! <loc> -iagra$ of 4:4 +it #edic *ulti)lier other represents DS(. C. Vedic Multiplier for #"# bit Module Ste) 2: $epresent the parts of , as , M and , D , and parts of ( The 2x2 bit )edic multiplier module as sho%n in the bloc" as ( M and ( D . <o% represent , and ( as , M , D and ( M ( D diagram in Fig. 7 can be easily implemented by using four 5x5 respectively. bit )edic multiplier modules as discussed in the previous Ste) =: For , = (, %e have general format as sho%n in Fig. section. DetGs analy;eiplications,say ,C2x2, I , 7 ,mult B , 5 , F , > , 0 , 8 and (C ( I ( 7 (B( 5 ( F ( > ( 0 ( 8 . The output line for the multiplication result %ill be of 07 bits as HS 0B S 05 S 0F S 0> S 00 S 08 S 1 S 2 S I S 7 S B S 5 S F S > S 0 S 8. DetGs divide t%o parts, say the 2 bit multiplicand , can be decomposed into pair of 5 bits ,6,D. Similarly multiplicand ( can be I decomposed into (6(D. The 07 bit product can be %ritten asM 8ig0 7 ?eneral 1e)re'entation for 77 International Journal of Engineering and Advanced Technology (IJEAT) ISSN: 2249 9!" #olu$e%&" I''ue%!" June 2(&2 P C , x ( C +,6,D- x +(6(D- C , x ( J +, x (D J ,D x (- J ,D x (D @sing the fundamental of )edic multiplication, ta"ing four bits at a time and using 5 bit multiplier bloc" as discussed %e can perform the multiplication. The outputs of 5x5 bit multipliers are added accordingly to obtain the final product. ere total three 2 bit $ipple6Carry ,dders are re&uired as sho%n in Fig. 7. I$)le$entation of an Efficient *ulti)lier +a'ed on #edic *athe$atic' u'ing E-A Tool #edic *ulti)lication Ste) 4: The individual multiplications product can be obtained by the partitioning method and applying the basic building bloc"s. (y adopting the above generali;ed algorithm %e can implement )edic Multiplier for any number of bits say 07, F>, 75, and so on, as per the re&uirement. Therefore, it could be possible to implement this )edic multiplier in the ,D@ +,rithmetic Dogic @nit- %hich %ill reduce the computational speed drastically / hence improves the processors efficiency. I#0 I*.4E*ENTATI2N IN @I4IN@ &20& 'n this %or", 2x2 bit )edic multiplier is designed in )DD +)ery igh Speed 'ntegrated Circuits ard%are Description Danguage-. Dogic synthesis and simulation %as done using *D, +*lectronic Design ,utomation- tool in =ilinx'S*0>.0i 6 Pro#ect <avigator and 'Sim simulator integrated in the =ilinx pac"age. The performance of circuit is evaluated on the =ilinx device family SpartanF, pac"age t&055 and speed grade 6B. The $TD schematic of 2x2 bit )edic multiplier ?vedicNmulticomprisesstruct2x2Aoffou r5x5bit)edic multipliervedicstruct5x5?N0A6 vm0, vm>, vmF, vm5 and three I6bit $ipple Carry ,dder ?rcNadderN2A6r0, r>, rF as sho%n in Fig. 2 %hile the simulation results obtained are sho%n in Fig. 1 for verification. 'n behavioral simulation %e have tested for the follo%ing input bitsM 6 8ig0 9 Si$ulation 1 e' ul t of :
+i t #e di c * ul ti )l ie r (! :; A= ( B : =A 24 ) #0 1ES,4T AN- -IS3,SSI 2N The synthesis result obtained from proposed )edic multiplier is faster than ,rray and (ooth multiplier. The device utili;ation summary of 2x2 bit )edic multiplier for =ilinx, Spartan family is sho%n belo%M -evice ,tiliCation Su$$ary: Selected DeviceM FsB8at &0556B <umber of SlicesM 1B out of I85 0FO <umber of 5 input D@TsM 077 out of 0582 00O <umber of '!sM F> <umber of bonded '!(sM F> out of 082 >1O a- For 2x2 bit )edic multiplier input, the multiplier aC?88888080A decimal+ number system B- and multiplicand bC?88888008A +decimal number076 bitcomputationalsystem pathdelay7-in nanosecondsand%e+ns-. Thegetpath delay output0C ?8888888888800008A +decimal-.for2x2bit,rr aynumberand(oothmul tiplierssystemhavebeen F8ta"en from S.S. 4erur et al. 900:. The timing result sho%s that )edic b- ,gain, %e have multiplier aC88880888A multiplier+decimalhasth egreatest advantage as compared to other n u m b e r s y s t e m
2 - a n d
m u l t i p l i c a n d
b C ? 8 8 8 8 8 0 0 8 A
m u l t i p l i e r s + d e c i m a l i n t e r m s o f
e x e c u t i o n
t i m e .
n u m b e r
s y s t e m
F -
% i t h
0 7 6 b i t
o u t p u t 0 C ? 8 8 8 8 8 8 8 8 8 8 8 0 0 8 8 8 A +decimal number system >5-. Ta+le & 3o$)ari'on of : +it *ulti)lier' (in n') -evice: S)artan Array <ooth :c='!(a%!%tD&44 *ulti)lier *ulti)lier Path Delay F>.808 ns >1.B51 ns #I0 32N34,SI2N This paper presents a highly efficient method of multiplication H ?@rdhva Tirya"bhyam SutraAbased on )edic mathematics. 't gives us method for hierarchical multiplier design and clearly indicates the computational advantages offered by )edic methods. The computational path delay for proposed 2x2 bit )edic multiplier is found to be >0.7I1 ns. ence our motivation to reduce delay is finely fulfilled. Therefore, %e observed that the )edic multiplier is much more efficient than ,rray and (ooth multiplier in terms of execution time +speed-. ,n a%areness of )edic mathematics can be effectively increased if it is included in engineering education. 'n future, all the ma#or universities may set up appropriate research centers to promote research 8ig0 1T4 'che$atic of : +it #edic *ulti)lier %or"s in )edic mathematics. 7 International Journal of Engineering and Advanced Technology (IJEAT) ISSN: 2249 9!" #olu$e%&" I''ue%!" June 2(&2 1E8E1EN3ES [1] 3agadguru S%ami, Sri (harati 4risna, Tirtha#i Mahara#a, ?)edic Mathematics or Sixteen Simple Mathematical Formulae from the )eda, Delhi,Motilal (anarsidas,+017B-A)aranasi, 'ndia, 0127. [2] M. Morris Mano, ?Computer System ,rchitectureA, F rd edition, Prientice6all, <e% 3ersey, @S,, 011F, pp. F576F52. [3] . Thapliyal and 6,rea.$6Po%er,rbania*fficient. ?, Time Multiplier and S&uare ,rchitecture (ased !n ,ncient 'ndian )edic MathematicsA, Proceedings of the >885 'nternational Conference on )DS' +)DS'G85-,Das)egas,<evada, 3une >885, pp. 5F565F1. [4] P. D. Chidgup"ar and M. T. 4arad, ?The 'mplementation of )edic ,lgorithms in Digital Signal*ngg.*du, ProcessingA, Plobal 3. of )ol.2, <o.>, >885, @'C** Published in ,ustralia. [5] Thapliyal . andigh SpeedSrinivas*fficient<x< (itM .(, ? Parallel ierarchical !verlay Multiplier ,rchitecture (ased on ,ncient 'ndian )edic MathematicsA, Transactions on *ngineering, Computing and Technology, >885, )ol.>. [6] arpreet Singh Dhillon andH(it ,bhi#it Mitra, ?, $educed Multipliction ,lgorithm for Digital ,rithmeticsA, 'nternational 3ournal of Computational and Mathematical Sciences >.> Q %%%.%aset.orgSpring>882. [7] oney Durga Ti%ari, Pan;orig Pan"huyag, Chan Mo 4im and Rong (eom Cho,Multiplier design? based on ancient 'ndian )edic MathematicianA, 'nternationalCDesignConference, ppSo.7B6 72, >882. [8] Parth Mehta and Dhanashri Pa%ali, ?Conventional versus )edic mathematics method for ard%are implementation of a multiplierA, 'nternational conference on ,dvances in Computing, Control, and Telecommunication Technologies, pp. 758675>, >881. [9] $amalatha, M.Dayalan, 4 D Dharani, P Priya, and S Deoborah, ?igh Speed *nergy *fficient ,D@ Design using )edic Multiplication Techni&uesA, 'nternational Conference on ,dvances in Computational Tools for *ngineering ,pplications +,CT*,- '***, pp. 788678F, 3uly 0B60I, >881. [10] Sumita )aidya and Deepa"6Po%erPerformanceDande"ar, ?Delay comparison of Multipliers in )DS' Circuit DesignA, 'nternational 3ournal of Computer <et%or"s / Communications +'3C<C-, )ol.>, <o.5, pp. 5I6B7, 3uly >808. [11] S.S.4erur, Pra"ash <archi, 3ayashree C <, arish M 4ittur and Pirish ) , ?'mplementation of )edicnal Multiplier For Digital Sig ProcessingA 'nternational conference on )DS' communication / instrumentation +'C)C'-, >800. [12] ,smita aveliya, ?, <ovel Design for igh Speed Multiplier for Digital Signal Processing ,pplications +,ncient 'ndian )edic mathematics'nternationalapproach-A,3ournalofTechnology and *ngineering System +'3T*S-, )ol.>, <o.0, pp. >I6F0, 3an6March, >800. [13] Prabha S., 4asli%al, (.P. Patil and D.4. Pautam, ?Performance *valuation of S&uaring !peration by )edic MathematicsA, '*T* 3ournal of $esearch, vol.BI, 'ssue 0, 3an6Feb >800. [14] ,niruddha 4anhe, Shishir 4umar Das and ,n"it 4umar Singh, ?Design and 'mplementation of Do% Po%er Multiplier @sing )edic Multiplication Techni&ueA,+'3CSC- 'nternational 3ournal of Computer Science and Communication )ol. F, <o. 0, pp. 0F060F>, 3anuary63une >80>. [15] @mesh ,"are, T.). More and $.S. Don"ar, ?Performance *valuation and Synthesis of )edic MultiplierA, <ational Conference on 'nnovative Paradigms in *ngineering / Technology +<C'P*T6>80>-, Proceedings published by 'nternational 3ournal of Computer ,pplications +'3C,-, pp. >86>F, >80>. *r'0 .u'h)alata #er$a obtained her (. Sc +Computer Science- in >88B from (.M.M., (hilai and M. Sc. +Computer Science- in >88I from S.S.M.)., (hilai. She is pursuing her M.*. in Computer Technology and ,pplication +CT,- from Shri Shan"aracharya College of *ngineering / Technology +SSC*T- (hilai +C.P.- 'ndia. Presently she is %or"ing as ,ssistant Professor in Central College of *ngineering / Technology +CC*M-, $aipur +C.P.- 'ndia. -r0 /0 /0 *ehta obtained his (.*. +Computers- in 0115 from 4'TS $amte" and M.Tech +Computers- from P*C $aipur in >88>. e is Ph.D. in Computers by Technical @niversity of Chhattisgarh State. Presently he is %or"ing as Professor / Dept ead at Shri Shan"aracharya College of *ngineering / Technology +SSC*T- (hilai +C.P.- 'ndia. is research area includes Do% Po%er Micro *lectronics, (us *ncoding Scheme, Cryptography and (iometric based application development. e is a life member of '*', 'ST*, CS' and )S' '<D',. 7 9