Improved RNS FIR filter architectures

被引:20
作者
Conway, R [1 ]
Nelson, J [1 ]
机构
[1] Univ Limerick, Elect & Comp Engn Dept, Limerick, Ireland
关键词
filter; residue number system (RNS);
D O I
10.1109/tcsii.2003.821524
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
A new architecture for implementing finite-impulse response (FIR) filters using the residue number system (RNS) is detailed. The design is based on using a restricted modulus set, with moduli of the form 2(n), 2(n) - 1, and 2(n) + 1. This does not restrict the modulus set to the common 3 modulus set {2(n) -1, 2(n), 2(n)+1}, but any number of pairwise relatively prime moduli of this form, for example, {5, 7, 17,31,32,33}. Based on a comparison with a 2's complement design, the new RNS design can offer a significant speed improvement. The gain is obtained by using a set of small moduli, selected so as to minimize critical path delay and area. An algorithmic approach is used to obtain full adder based architectures that are optimized for area and delay. The modulus set is optimum based on cost parameters for each modulus. This new architecture presents a practical approach to implementing a fast RNS FIR filter.
引用
收藏
页码:26 / 28
页数:3
相关论文
共 10 条
[1]   A NOTE ON FREE ACCUMULATION IN VLSI FILTER ARCHITECTURES [J].
CAPPELLO, PR ;
STEIGLITZ, K .
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS, 1985, 32 (03) :291-296
[2]   New CRT-Based RNS converter using restricted moduli set [J].
Conway, R ;
Nelson, J .
IEEE TRANSACTIONS ON COMPUTERS, 2003, 52 (05) :572-578
[3]  
CONWAY R, 2001, THESIS U LIMERICK
[4]  
MOHAN A, 2000, RESIDUE NUMBER SYSTE
[5]   A method for speed optimized partial product reduction and generation of fast parallel multipliers using an algorithmic approach [J].
Oklobdzija, VG ;
Villeger, D ;
Liu, SS .
IEEE TRANSACTIONS ON COMPUTERS, 1996, 45 (03) :294-306
[6]  
SWARTZLANDER EE, 1978, P 4 IEEE S COMP AR O
[7]  
WANG Z, 1996, J VLSI SIGNAL PROC, V14, P242
[8]  
Wei SG, 1996, IEICE T INF SYST, VE79D, P242
[9]  
WESTE NHE, 1993, PRINCIPLES CMOS VLSI
[10]  
WRZYSZCZ A, 1993, IEEE INT C COMP DES, P614