A unified and reconfigurable Montgomery Multiplier architecture without four-to-two CSA

Authors:
Sudhakar Maddi;M. B. Srinivas
Affiliations:
International Institute of Information Technology (IIIT), Hyderabad, India;International Institute of Information Technology (IIIT), Hyderabad, India
Venue:
Proceedings of the 20th annual conference on Integrated circuits and systems design
Year:
2007

Citing 7
Cited 1

Network and internetwork security: principles and practice

Network and internetwork security: principles and practice
A method for obtaining digital signatures and public-key cryptosystems

Communications of the ACM
Handbook of Applied Cryptography

Handbook of Applied Cryptography
Elliptic Curve Public Key Cryptosystems

Elliptic Curve Public Key Cryptosystems
Montgomery Modular Exponentiation on Reconfigurable Hardware

ARITH '99 Proceedings of the 14th IEEE Symposium on Computer Arithmetic
A Scalable Architecture for Modular Multiplication Based on Montgomery's Algorithm

IEEE Transactions on Computers
An Improved Unified Scalable Radix-2 Montgomery Multiplier

ARITH '05 Proceedings of the 17th IEEE Symposium on Computer Arithmetic

Area/performance trade-off analysis of an FPGA digit-serial GF(2m) Montgomery multiplier based on LFSR

Computers and Electrical Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an improved version of unified Montgomery multiplier architecture to reduce the critical path delay leading to higher data throughput. It is known that the critical path delay due to four-to-two carry save adders (CSAs), which have two levels of carry-save logic, is predominant in determining the overall performance of the CSA based multipliers. In this paper, authors improve the multiplier efficiency by replacing four-to-two CSA with new one level area-efficient sum-carry logic. While conventional four-to-two CSA has a delay of four .xor. gates, the proposed sum-carry logic has a delay of only two .xor. gates. Not only does this result in reduced critical path delay but also results in reduced area compared to previous multiplier architectures. Also it can compute n-bit multiplication in n+2 clock cycles in both fields GF(p) and GF(2n). Further, the critical path delay is independent of input operand precision.