Article in HTML

Author(s): Anmol Nagar, Sheetal Nagar

Email(s): anmol319nagar@gmail.com

Address:

    Department of Electronics and Communication Engineering, IIMT College of Engineering, Greater Noida, UP, India

Published In:   Volume - 5,      Issue - 1,     Year - 2025


Cite this article:
Anmol Nagar, Sheetal Nagar (2025), Design and Optimization of Booth Multipliers for High-Speed Digital Arithmetic, Spectrum of Emerging Sciences, 5 (1) 46-50, 10.55878/SES2025-5-1-9

  View PDF

Please allow Pop-Up for this website to view PDF file.



I INTRODUCTION

Multiplication is a essential operation in digital signal processing, computer mathematics, and other numerous computing packages. traditional multiplication algorithms are normally sluggish and inefficient for big-bit-width operations. A powerful approach was developed to enhance the efficiency of multiplication by way of Andrew D. booth in 1951 referred to as booth's multiplication algorithm[1].

Booth's set of rules achieves this by means of encoding the multiplier using a way known as sales

space recoding, which correctly minimizes the range of required arithmetic operations, especially while coping with signed numbers.[2] This makes the sales space multiplier specifically nice for hardware implementations, consisting of in virtual processors and FPGA-based designs, wherein velocity and strength performance are vital[3]. The booth multiplier is broadly utilized in present day computing structures because of its capability to address both high quality and bad operands efficaciously even as enhancing performance over conventional shift-and add multiplication strategies. This paper explores the principles, implementation, and optimizations of the sales space multiplier, highlighting its advantages, boundaries, and capacity applications in modern virtual systems.[4]

 

The primary gain of the booth multiplier is its capacity to accelerate multiplication for numbers with long sequences of 1s or 0s. This makes it fantastically beneficial for hardware implementations which includes arithmetic logic devices (ALUs), digital sign Processors (DSPs), and subject Programmable Gate                   Arrays (FPGAs). In addition, sales space's multiplication is carried out in excessive-performance computing packages in which optimization of computational velocity and resource usage is of incredible significance.[5]

Even though a green set of rules, sales space's has positive drawbacks. It reduces the number of operations for specific forms of numbers worried, but it introduces overhead while more bit manipulations are wanted. To acquire those, two improved variations of booth's set of rules have come to be well known: changed sales space algorithm (MBA) and Radix-four sales space Encoding. [6]

 

II WORKING ALGORITHM

Booth’s Multiplication algorithm is an effective technique used for binary multiplication, mainly for signed numbers in two’s complement shape. It minimizes the number of arithmetic operations via encoding the multiplier using booth recoding, which reduces the wide variety of required partial products. This consequences in a faster and more hardware-green multiplication method.[7]

1.  Precept of Booth’s algorithm

Booth’s algorithm works with the aid of inspecting the bits of the multiplier in pairs and figuring out whether or not to add, subtract, or shift primarily based at the bit styles. The key remark in booth’s recoding is that sequences of 1s in a binary range may be replaced with fewer mathematics operations, lowering the computational complexity.[8]

Rather than performing a honest shift-and-upload multiplication, booth’s set of rules follows a recoding rule based totally on two adjacent bits of the multiplier:

If 01 is encountered, the multiplicand is delivered to the product.

If 10 is encountered, the multiplicand is subtracted from the product.

If 00 or eleven is encountered, best a right shift is achieved (no addition/subtraction).

This pattern is recognized through appending an extra least large bit (LSB) of the multiplier, referred to as Q-1, initialized to 0.[9]

2.  Steps of Booth’s Multiplication set of rules

Bear in mind multiplying n-bit numbers, in which M is the multiplicand and Q is the multiplier. The set of rules follows those steps:

 

   Fig 2: Output of Simulation of Booth multiplier code in Xilinx ISE

 Initialize Registers:

 

The Multiplicand (M) is stored in a register. The Multiplier (Q) is saved in another register.

a further bit Q-1 is appended to Q, initialized to 0.

The Accumulator (A) is initialized to 0 (holds the intermediate and very last product).

The quantity of iterations required is same to the range of bits inside the multiplier.

look at the last two bits (Q₀ and Q₋₁):

 

If Q₀Q₋₁ = 01, add M to the accumulator (A = A + M).

If Q₀Q₋₁ = 10, subtract M from the accumulator (A = A - M).

If Q₀Q₋₁ = 00 or 11, do nothing (handiest shift). carry out an arithmetic proper Shift (ARS):

Shift A, Q, and Q-1 right by means of one bit.

preserve signal extension to maintain the wide variety’s register ’s supplement form.

Repeat steps 2 and 3 for n iterations (wherein n is the range of bits in the multiplier).

 

The final end result is stored in the concatenated A and Q registers.[6]

 

III. WORKING EXAMPLE

Let’s multiply (-6 × 3) using Booth’s Algorithm, assuming 4-bit binary representation.

 

M (Multiplicand) = -6 1010 (in 2’s complement)

Q (Multiplier) = 3 0011

Q-1 = 0 (initialized to zero)

A = 0000 (initially zero)


               


Thus, (-6 × 3) = -18, which is correctly obtained using Booth’s Algorithm.

 

DIAGRAM

The block diagram represents the Booth Multiplier, a hardware implementation of Booth’s algorithm for signed binary multiplication. Let’s break down its functioning part by part:


Fig. 1: Block Diagram of Booth Multiplier

 

Multiplier Buffer

The Multiplier (one of the operands) (For example (0.7)) is first stored in a buffer.

Booth Encoder

It examines groups of bits (usually two at a time) and generates control signals for multiplication.

Multiplicand Buffer

The Multiplicand (the second operand) (Taking (0.7) as example value) is stored in another buffer.

Partial Product Generator

It utilizes shift and add/subtract operations to form intermediate results.

Adder (Final Stage)

The partial products generated in the previous stage are summed using an adder. Then the final product is received as (15:0).

 

OUTPUT

 

Fig 2: Output of Simulation of Booth multiplier code in Xilinx ISE

 

 

Fig. 3: Circuit Diagram Output of Booth Multiplier code in Xilinx ISE

 

 

Fig. 4: Testing example of Booth Multiplier code in Xilinx ISE

 

Figure 2 shows the output of initial simulated code of Booth multiplier with initial inputs of (7:0) and getting (15:0) as an output in Xilinx ISE 14.7. Figure 3 shows the output circuit of booth multiplier in Xilinx ISE 14.7 and Figure 4 shows an tested example of booth multiplier in Xilinx ISE 14.7 where initial values were changed from (7:0) to (3:0) and the final product we got was changed from (15:0) to (7:0).

 

ADVANTAGES OF BOOTH MULTIPLIER

 

Reduces the number of arithmetic operations, improving multiplication efficiency.

Handles signed number multiplication easily without extra sign adjustments.

Useful for hardware implementations, such as in ALUs, DSPs, and FPGAs.

Works efficiently for multipliers with long sequences of 1s, reducing operations.

 

LIMITATIONS OF BOOTH MULTIPLIER

 

If the multiplier does not have long sequences of 1s, the performance improvement is limited.

The arithmetic right shift requires sign extension, which can increase hardware complexity.

Extra steps are required for handling negative numbers in two’s complement form.

 

POSSIBLE APPLICATION AREAS

 

Digital Signal Processing (DSP) Systems: Booth multipliers are crucial in various DSP applications such as Fast Fourier Transform (FFT) algorithms, digital filters (like FIR and IIR), and modulation/demodulation techniques used in communication systems. Their speed and efficiency are critical for real-time signal processing demands.

 

Graphics Processing Units (GPUs): Modern GPUs heavily rely on high-performance arithmetic units, including multipliers. Booth multipliers are employed in the design of GPU cores to accelerate operations like matrix multiplication, texture mapping, and 3D rendering.

 

Digital Communication Systems: In communication systems, Booth multipliers are essential for tasks like modulation (e.g., QAM, PSK) and demodulation, where complex multiplications are frequently required. Their speed is crucial for high-bandwidth communication systems.

 

Cryptography: Cryptographic algorithms often involve numerous multiplication operations. Booth multipliers can be used to accelerate the computation of cryptographic primitives like elliptic curve cryptography (ECC) and RSA, enhancing the speed and efficiency of security systems.

 

Machine Learning Accelerators: Deep learning models involve extensive matrix multiplications. Dedicated hardware accelerators, often employing Booth multipliers, can significantly speed up training and inference processes for neural networks, enabling faster and more energy-efficient AI applications

 

IV. CONCLUSION:

Booth’s Multiplication Algorithm is a widely used technique for efficient binary multiplication, particularly for signed numbers. By reducing the number of partial products, it enhances computation speed and simplifies hardware implementation. Despite some limitations, Booth’s algorithm remains a fundamental method in digital arithmetic and is further improved by modified versions for even greater efficiency.



Related Images:

Recomonded Articles:

Author(s): Juhi Mishra; Sapna Sorrot; Seema Nayak; Puneet Mittal

DOI: 10.55878/SES2024-4-1-7         Access: Open Access Read More

Author(s): Punit Tomar, Ankit Sharma, Sandhya Bhardwaj

DOI: 10.55878/SES2025-5-1-3         Access: Open Access Read More

Author(s): Achitya Srivastava, Arpit Dubey, Dev Prakash, Surendra Kumar

DOI: 10.55878/SES2025-5-1-5         Access: Open Access Read More

Author(s): Anmol Nagar, Sheetal Nagar

DOI: 10.55878/SES2025-5-1-9         Access: Open Access Read More

Author(s): Ram Ashish Maurya, Riya Tiwari, Aayush Vikram Singh

DOI: 10.55878/SES2025-5-2-7         Access: Open Access Read More

Author(s): Rishabh Raj, Ritesh Kumar, Shubham Kumar

DOI: 10.55878/SES2025-5-2-9         Access: Open Access Read More

Author(s): Anush Kumar Singh, Ankit Kumar, Surendra Kumar

DOI: 10.55878/SES2025-5-2-14         Access: Open Access Read More

Author(s): Akash Tiwari, Amar Kishor, Surendra Kumar

DOI: 10.55878/SES2025-5-2-16         Access: Open Access Read More

Author(s): Aniket Pandey, Mohd. Suleman Khan, Km. Shaban Ahmad, Rishabh kumar, Danish Nayab, Saumitra Pal

DOI: 10.55878/SES2022-2-1-14         Access: Open Access Read More

Author(s): Kuldeep, Sanjeev Kumar, Nikhil Kumar, Deepak Yadav, Mohd. Shahbaz, Shailendra Vikram Yadav, Greeshma Srivastava

DOI: 10.55878/SES2024-4-1-21         Access: Open Access Read More

Author(s): Aman Kumar, Gaurav Rai, Satyam Maurya, Dileep Kumar Singh, Greeshma Srivastava

DOI: 10.55878/SES2024-4-1-18         Access: Open Access Read More

Author(s): Amit Yadav, Samrendra Singh, Abdul Fahad

DOI: 10.55878/SES2023-3-1-13         Access: Open Access Read More

Author(s): Samrendra Singh, Aditya Pathak, Saurabh Kumar, Svostik Kumar, Vinay Kumar Yadav, Zeeshan Vakil

DOI: 10.55878/SES2023-3-1-11         Access: Open Access Read More

Author(s): Mohd. Ahsan, Surender Kumar, Abhishek Pandey, Ayush Singh, Himanshu Pathak, Dibya Prakash, Kaustubh Kundan Srivastava

DOI: 10.55878/SES2024-4-1-20         Access: Open Access Read More