Performance Analysis of FPGA Based MAC Unit using DBTNS Multiplier & TRNS Adder for Signal Processing Algorithm

Digital signal processing (DSP) algorithms are actually nothing but sum of product. So, they are computationally intensive. All the mathematical tasks related to DSP algorithm are based on multiplication and addition. So, the implementations of DSP algorithms-based applications extensively require multiplier and adder. Due to extensive use of multiplication and addition operation, speed up cannot be achieved. Designing the high-performance adder and multiplier are primary objective for implementing high performance signal processing applications. This hindrance can be removed by Multiply-Accumulate Unit (MAC). The special feature of a MAC unit is its ability to perform single cycle multiplication and addition operation.The performance of MAC Unit can be improved by using non-binary number system. Ternary value logic (TVL) has the ability to offer several advantages over conventional binary number system like reduce chip area, reduce overall delay. TVL can switch between three levels, such levels denoted by 0, 1, 2. As Residue Number Systems (RNS) can perform "carry free" arithmetic operations, high performance adder can be implemented using RNS in TVL domain (TRNS).  Partial product free multiplication can be implemented by using Double Base Number System in TVL domain (DBTNS). Double Base Ternary Number System (DBTNS) multiplier can perform better than conventional TVL multiplier. A MAC unit is used to perform the multiplication and accumulator operations together to avoid unnecessary overhead on the processor in terms of processing time and the on-chip memory requirements. Keeping in view of these issues, a new architecture is proposed for implementing high performance MAC unit for DSP applications. In this paper, a new approach of designing MAC unit is instigated using DBTNS multiplier and TRNS adder.  A major bottleneck of implementing this architecture is the complexity involved in converting TVL to DBTNS in initial stage and converting TRNS to TVL in final stage. The performance of Ternary Residue Number Systems (TRNS) system depends on selection of moduli because selection of moduli is not properly maintained then it will affect system speed, dynamic range and hardware complexity. Proposed MAC unit is mapped on field programmable gate array (FPGA) for analysis its performance.

[1]  Reto Zimmermann.  Computer Arithmetic: Principles, Architectures, and VLSI Design , 1999 .

[2]  B. Lee.  A new algorithm to compute the discrete cosine Transform , 1984 .

[4]  B. Radanović,et al.  Current-mode CMOS adders using multiple-valued logic , 1996, Proceedings of 1996 Canadian Conference on Electrical and Computer Engineering.

[5]  Graham A. Jullien,et al.  Near canonic double-based number system (DBNS) with applications in digital signal processing , 1996, Optics & Photonics.

[6]  Thanos Stouraitis,et al.  Efficient analog-to-residue conversion schemes , 1990, IEEE International Symposium on Circuits and Systems.

[7]  Russell Tessier,et al.  c ○ 2001 Kluwer Academic Publishers. Manufactured in The Netherlands. Reconfigurable Computing for Digital Signal Processing: A Survey ∗ , 1999 .

[8]  Behrooz Parhami,et al.  Computer arithmetic - algorithms and hardware designs , 1999 .

[9]  Majid Ahmadi,et al.  A hybrid DBNS processor for DSP computation , 1999, ISCAS'99. Proceedings of the 1999 IEEE International Symposium on Circuits and Systems VLSI (Cat. No.99CH36349).

[10]  Nevio Benvenuto,et al.  Realization of Finite Impulse Response Filters Using Coefficients +1, 0, and -1 , 1985, IEEE Trans. Commun..

[11]  Roberto Saletti,et al.  A New CMOS Ternary Logic Design for Low-power Low-voltage Circuits , 2000 .

[12]  Amitabha Sinha,et al.  A new architecture for FPGA implementation of a MAC unit for digital signal processors using mixed number system , 2012, CARN.

[13]  F. J. Taylor,et al.  Residue Arithmetic A Tutorial with Examples , 1984, Computer.

[14]  W.C. Miller,et al.  On efficient techniques for difficult operations in one and two-digit DBNS index calculus , 2000, Conference Record of the Thirty-Fourth Asilomar Conference on Signals, Systems and Computers (Cat. No.00CH37154).

[15]  Keivan Navi,et al.  A New Moduli Set for Residue Number System in Ternary Valued Logic , 2007 .

[16]  Roberto Muscedere,et al.  A 2-digit DBNS filter architecture , 2000, 2000 IEEE Workshop on SiGNAL PROCESSING SYSTEMS. SiPS 2000. Design and Implementation (Cat. No.00TH8528).

[17]  Ping Wah Wong.  Fully sigma-delta modulation encoded FIR filters , 1992, IEEE Trans. Signal Process..

[18]  Amitabha Sinha,et al.  FPGA Implementation of MAC Unit for Double Base Ternary Number System (DBTNS) and its Performance Analysis , 2018 .

[19]  Chung-Yu Wu,et al.  Design and application of pipelined dynamic CMOS ternary logic and simple ternary differential logic , 1993 .

[20]  M. Yoeli,et al.  Logical Design of Ternary Switching Circuits , 1965, IEEE Trans. Electron. Comput..

[21]  Amitabha Sinha,et al.  A new architecture for FPGA based implementation of conversion of binary to double base number system (DBNS) using parallel search technique , 2012, CARN.

[22]  Ronald Smith,et al.  Computer Organization , 1965, SIGCPR '65.

[23]  Graham A. Jullien,et al.  Theory and applications for a double-base number system , 1997, Proceedings 13th IEEE Sympsoium on Computer Arithmetic.

[24]  P. Siy,et al.  A new moduli set selection technique to improve sign detection and number comparison in residue number system (RNS) , 2005, NAFIPS 2005 - 2005 Annual Meeting of the North American Fuzzy Information Processing Society.

[25]  M. Omair Ahmad,et al.  Moduli selection in RNS for efficient VLSI implementation , 2003, Proceedings of the 2003 International Symposium on Circuits and Systems, 2003. ISCAS '03..

[26]  Richard Conway,et al.  Improved RNS FIR filter architectures , 2004, IEEE Transactions on Circuits and Systems II: Express Briefs.

[27]  Pinaki Mazumder,et al.  Multiple-Valued Signed-Digit Adder Using Negative Differential-Resistance Devices , 1998, IEEE Trans. Computers.