analysis of 8 x 8 bit

Upload: dwarakamaneekantaaddagiri

Post on 25-Feb-2018

216 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/25/2019 Analysis of 8 X 8 Bit

    1/4

    Comparative Analysis of 8 X 8 Bit

    Vedic and Booth Multiplier

    Sasha Garg, Swati Garg, Vidhi SachdevaElectronics and Communication

    ITM UniversityGurgaon, INDIA

    [email protected], [email protected],[email protected]

    Abstract - Speed and power consumption are one of the mostimportant parameters to judge the performance of a

    computational method. In this paper, we compare two

    algorithms for 8 Bit multiplication namely Vedic

    Multiplication Algorithm and Booth algorithm. This paper

    aims in bringing to the fore the differences in compilation

    speeds and the chip area consumption of the two

    methodologies. The programming language used is Verilog andthe synthesis has been done on Xilinx 14.5.

    Keywords: Vedic Multiplication, Booth Multiplication, Xilinx,

    Verilog

    I. INTRODUCTIONMultiplication forms one of the basic arithmetic

    operations that find itself in wide use, especially in DigitalSignal Processing. A variety of computerarithmetic techniques are available which can be used inimplementation of a digital multiplier. Most of thesetechniques employ the concept of a set of partial productsand their simultaneous addition. The above process issimilar to the method that is taught to schoolchildren forconducting long multiplication on decimals, but has beenmodified here for application to binary digits. Variousmultiplication techniques are available in the field ofmathematics like, repetitive addition, Vedic Multiplication,Booth Multiplication, Ancient Egyptian Method etc. Thecomparison is based on 3 parameters: delay, area, power.These form the basis for judging the efficiency of the VLSIdesign of the multiplier in concern. The ancient technique ofVedic multiplication involves the use of UrdhavaTiryagbhyam Sutra[2]. Booth Multiplication algorithm[6,7]uses recorded multipliers. Vedic mathematics is mainly

    based on sixteen principles or word-formulae which are

    termed as sutras one of which is Urdhava TiryagbhyamSutra. Booth's multiplication algorithm is a multiplicationalgorithmthat multiplies two signed binary numbersin twos complements notation. In this paper, we focus on 8X 8 Bit multiplier using the above two mentionedtechniques implemented and simulated on Xilinx Version14.5.

    II. BOOTH MULTIPLIERIn Booth multiplication the compilation speed it

    increased due to the decrease in the number of partialproducts required to compute the result[1]. The algorithmmakes use of simple operations like twos complement,arithmetic right shift etc.; instead of more complicated ones

    like addition. The architecture takes two 8 bit numbers asinput and generates a 16 bit product as the output. Since, weare performing signed multiplication the MSB of theoperands is always the signed bit and will not be used as

    part of the magnitude. To decide on the multiplier and themultiplicand, we take the 2s complement of the twonumbers. The numerically smaller number is designated to

    be the multiplicand and the other to be the multiplier. Nowthe beginning product is obtained by concatenating 8 0-bitswith the multiplier. The previous LSB is initialized as 0.Thecurrent LSB and the previous LSB together decide thecourse of the arithmetic action to be taken according to thefollowing:

    00: no arithmetic operation

    01: add multiplicand to left half of the product

    10: subtract multiplicand from left half of theproduct

    11: no arithmetic operation

    Next, an arithmetic right shift (ASR) is performed on theentire product. The above process is repeated 8 times (sizeof operands).At the end of the 8

    thpass, the final product is

    obtained.

    For hardware implementation of multiplication, same steps

    are replicated as is done by hand, i.e., partial products are

    computed, shifted accordingly, and summed. This algorithmthus, becomes very slow if there are many partial products

    (i.e. many bits) as the output needs to wait until each sum isperformed. Booths algorithm decreases the number of

    required partial products and hence is faster owing to thefact that the number of sums of partial products wasreduced.

    2607978-1-4799-3080-7/14/$31.00 c2014 IEEE

  • 7/25/2019 Analysis of 8 X 8 Bit

    2/4

    III. VEDIC MULTIPLIERThe Vedic Multiplier[2,3] uses the Urdhva Tiryagbyham

    Sutra. It is also known as Vertical and Cross Wise method.

    In this the partial products are obtained with the subsequentaddition of all the partial products.

    For example, take two numbers and place them onebelow the other.

    The following steps are then performed:

    Multiply the digits at ones place with each otherand write it as the right most part of the answer.

    Multiply the digits at tens place with each other.This is the left most part of the answer.

    Then multiply the numbers crosswise and add theproducts to get the middle part.

    If any carry is generated, it is added to the numberto its left.

    Hence the answer is obtained.

    In the case of Binary numbers, the multiplication isreplaced by AND logic. For a 4 bit multiplier consider twonumbers a3a2a1a0and b3b2b1b0. By using Vedic methods the

    result obtained is a number which is more than 4 bits. Henceit is taken to be R7R6R5R4R3R2R1R0. The result is obtainedto be[1]:

    i) R0= a0b0ii) R1= a0b1+ a1b0iii) R2= a0b2+ a1b1+ a2b0+ R1iv) R3 = a0b3+ a1b2+ a2b1+ a3b0+ R2v) R4= a1b3+ a2b2+ a3b1+ R3vi) R5= a1b2+ a2b1+ R4

    vii) R6= a3b3+ R5viii) R7= carry of R6

    An 8 bit multiplier is generated using four 4 bitmultipliers (which are formed with the help of AND gates,half adders and full adders) and 3 ripple carry adders. Theripple carry adders are formed by the combinations of XORgates, full adders and half adders. The ADDER 1 consists of7 full adders and a half adder. ADDER 2 comprises of fourhalf adders, three full adders and a XOR gate. ADDER 3has three half adders, four full adders and a XOR gate[4,5].

    ARITHMETIC SHIFTRIGHTp,Q,cy_p

    COUNT

  • 7/25/2019 Analysis of 8 X 8 Bit

    3/4

    Fig. 3. OUTPUT FOR BOOTH MULTIPLIER

    Fig. 4. OUTPUT FOR VEDIC MULTIPLIER

    IV. COMPARISIONBased on our experimental findings of the comparison

    conducted on the Nexys3, Spartan 6, the differences

    between 8X8 Vedic Multiplier and Booth Multiplier can besummarized as follows:

    The area in terms of utilization of slices is comparativelyless in Vedic Multiplier. The Vedic Multiplier uses 161slices whereas the Booth Multiplier uses 287 slices. Thenumber of fully used LUT-FF pairs is 0 in case of VedicMultiplier and 1 in case of Booth Multiplier.

    The delay generated in Vedic Multiplier is approximately1/10th of the delay generated in Booth Multiplier.

    TABLE 1. DELAY SUMMARY (BOOTH)

    DELAY SUMMARY (BOOTH)

    Delay 73.962 ns (Levels of Logic = 186)

    Source u_7 (LATCH)

    Destination u_7 (LATCH)

    Source Clock x[7]_x[6]_Mux_o falling

    Destination Clock x[7]_x[6]_Mux_o falling

    TABLE 2. DELAY SUMMARY (VEDIC)

    DELAY SUMMARY (VEDIC)

    Delay 7.929 ns (Levels of Logic = 13)

    Source x (PAD)

    Destination Prod (PAD)

    2014 International Conference on Advances in Computing,Communications and Informatics (ICACCI) 2609

  • 7/25/2019 Analysis of 8 X 8 Bit

    4/4

    Fig. 9. COMPARISON (Delay)

    Fig. 10. COMPARISON (Area)

    V. CONCLUSIONThe Urdhva Tiryagbyham Sutra provides us with a

    highly efficient method over Booth Algorithm formultiplication in terms of speed, delay and area utilization.Therefore, it can be concluded that the 8X8 bit VedicMultiplier is better than Booth Multiplier based on the

    parameters discussed above. The specified architectures for

    Vedic Multiplication and Booth Multiplication can beextended for 16- bit, 64-bit etc. multiplication.

    VI. ACKNOWLEGMENTThis research paper is made possible through the help

    and support from our Department of EECE, ITM Universitywho provided us with the necessary software and hardwareincluding Xilinx and FPGA Kit.

    We would like to acknowledge the efforts and guidanceoffered by our faculty.

    VII. REFERENCES[1] Premananda B.S.,Samarth S. Pai, Shashank B, Shashank S.

    Bhat, Design and Implementation of 8-Bit Vedic Multiplier,

    IJAREEIE, Volume. 2, Issue 12, December 2013

    [2] Swami Bharati Krishna Tirthaji Maharaja, VedicMathematics, MotilalBanarsidass Publishers, 1965.

    [3] Ch. Harish Kumar "Implementation and Analysis of Power,Area and Delay of Array Urdhava, Nikhilam Vedic Multipliers," Internatinal Journal of Scientific and Research Publications,

    Volume 3, Issue 1 January 2013, ISSN 2250 3153.

    [4] Anju and V.K. Agrawal, FPGA Implementation of Low Power

    and High Speed Vedic Multiplier using Vedic Mathematics,IOSR Journal of VLSI and Signal Processing (IOSR-JVSP)Volume 2, Issue 5 Jun. 2013, ISSN: 2319 4200, pp. 51-57.

    [5] Aniruddha Kanhe, Shishir Kumar Dasand Ankit Kumar Singh,Design and Implementation of Low Power Multiplier usingVedic Multiplication Technique, International Journal of

    Computer Science and Communication, Volume 3, No. 1, June

    2012, pp. 131-132.[6] Dr. Ravi Shankar Mishra, Prof. Puran Gour and Braj Bihari

    Soni, 'Design and Implements of Booth and Robertsons

    multipliers algorithm on FPGA',International Journal ofEngineering Research and Applications (IJERA)Vol 1(Issue 3),

    pp. 905-910.

    [7] Neha Goyal, Khushboo Gupta and Renu Singla (2014), 'Studyof Combinational and Booth Multiplier',International Journalof Scientific and Research PublicationsVolume 4(Issue 5).

    .

    0

    20

    40

    60

    80

    Delay(inns)

    Vedic Mul.

    Booth Mul.

    0

    100

    200

    300

    400

    NumberofLUTSlices

    used Vedic Mul

    Booth Mul2

    2610 2014 International Conference on Advances in Computing,Communications and Informatics (ICACCI)