Energy Efficient Novel Design of Static Random Access Memory Memory Cell in Quantum-dot Cellular Automata Approach

S. Kassa and S. Nema / IJE TRANSACTIONS B: Applications Vol. 32, No. 5, (May 2019) 720-725 721 are the primary gates for building any QCA based logic circuits [6, 7]. Figure 1 shows the QCA cell, QCA inverter and QCA wire while Figure 2 shows the Original, Rotated and 5-input majority gate consequently. At the moment, there is a growing demand for memory cell design which works as one of the bottomline requirements for QCA devices [10–13]. In this article, loop-based approach have been taken for designing well-optimized SRAM cell compared to traditional CMOS based SRAM memory cell, which is a pioneering approach. Two main approaches are by and large available for designing QCA circuits: loop-based and line-based. In loop-based approach, all the clock zones are correlated to hold data in a loop of QCA cells while in line-based approach, a line of QCA cells is utilized for saving earlier values of the data. In the proposed design here, loopbased approach has been utilized for its simplicity and robustness in designing circuit. (a) (b) (C) Figure 1. QCA (a) Cell (b) Inverter (c) Wire (a) (b) (c) Figure 2. QCA based (a) Original (b) Rotated (c) 5-input Majority Gate 2. PROPOSED QCA SRAM CELL DESIGN Discussion about the robust and efficient structure of QCA SRAM cell has been preceded in this section. Figures 3 and 4 show the SRAM cell designs from previous papers. Figures 5a, 5b and 5c show logical diagram, layout and simulation result for the proposed QCA SRAM cell accordingly. 3. OPERATION OF PROPOSED SRAM CELL The proposed design configuration join one 3-input MG (MG-3) and one 5-input MG (MG-5) in addition to one (2×1) Multiplexer block as shown in Figure 5a. While the Enable signal is set to logic 1, the write operation can also be accomplished by setting W/R signal to logic 1. Along this line, the 3-input MG produces logic 1 since most of the inputs are at logic 1. Most of the time, one gets odd number of inputs, so they don’t have to deal with the question of what happens when exactly half the inputs are logic 0. During the write operation, the input data (Input), Figure 3. SRAM cell design as reported in [14] Figure 4. SRAM cell design as reported in [15] 722 S. Kassa and S. Nema / IJE TRANSACTIONS B: Applications Vol. 32, No. 5, (May 2019) 720-725 (a) (b) (c) Figure 5. QCA SRAM cell (a) logical diagram (b) layout (c) simulation result which is fed to the second input of the (2×1) Multiplexer (I1), will be written to the memory loop (MLoop) and also transmitted to the 5-input MG together with the Enable signal, two static logic 0 inputs and the inverted W/R signal (logic 0). The MG-5 with three static logic 0 inputs gives outputs as logic 0. It is worthy to take note of that, there is not any tri-state support idea in QCA innovation available, which totally eliminates the particular information in a circuit. Typically, logic 0 has been considered as the output of memory cell in write mode when output circuit is deactivated. The complete operation of the proposed SRAM cell is given in Table 1. When Enable input is ‘0’, the output will hold the previous output as it is for whatever input it would be. When Write operation is enabled, W/R should be ‘1’ and if Input is ‘1’, so ‘1’ will be written in MLoop and if Input is ‘0’, so ‘0’ will be written in MLoop. Like that, when Read operation is enabled, W/R should be ‘0’. Now, based on the signal in MLoop, Output will read the particular signal ‘0’/’1’ at an Output. 4. RESULT ANALYSIS Comparison of proposed SRAM cell design with previous designs is given in Table 2 for different parameters. Structural analysis of the proposed cell is verified utilizing QCADesigner 2.0.3 tool [16]. Table 3 shows the comparison of different types of power dissipation for the proposed and reported designs obtaining the results from QCAPro tool [17]. Figure 6 shows the power dissipation analysis of the proposed SRAM cell design with previous designs at different kink energies (Ek). The results show the dominance of the proposed design on the former reported designs. TABEL 1. Complete operation of proposed QCA SRAM Memory Cell Operation Enable Input W/R MLoop Output Write Read 1 × 1 × 0 1 1 Hold 0 × × Hold 0 TABLE 2. Comparison of the proposed QCA SRAM Circuit QCA SRAM cells Type of wire crossing Cell Count Area (μm2) Cycles Output (via W/R and Enable signals) Proposed Logical 92 0.10 1.5 YES Previous design [14] Coplanar 100 0.12 2 NO Previous design [15] Coplanar 109 0.13 2 NO S. Kassa and S. Nema / IJE TRANSACTIONS B: Applications Vol. 32, No. 5, (May 2019) 720-725 723 TABLE 3. SRAM Power dissipation analysis Ek=0.5 (meV) × 10-2 Ek=1 (meV) × 10-2 Ek=1.5 (meV) × 10-2 Previous design [14] Previous design [15] Proposed design Previous design [14] Previous design [15] Proposed design Previous design [14] Previous design [15] Proposed design Max Energy Diss of Circuit 44.41 42.55 28.69 47.69 43.93 30.97 53.25 46.82 34.42 Avg Energy Diss of Circuit 22.68 18.89 13.92 29.46 23.38 18.31 37.85 29.19 23.63 Min Energy Diss of Circuit 4.69 2.77 3.33 13.83 9.12 8.97 24.45 16.88 15.62 Avg Leakage Energy Diss 4.98 3.01 3.33 14.46 9.55 9.26 25.26 17.4 15.92 Avg Switching Energy Diss 17.69 15.89 10.59 14.99 13.83 9.05 12.59 11.79 7.71 4. 1. Sram Power Dissipation Analysis Figure 7 shows the power dissipation analysis for the proposed SRAM cell. The red color cell shows the lowest level of energy dissipation by a particular cell while the black color spot shows the highest level of energy dissipation by the cell. From the figure, one can analyze that, at a (a) (b) (c) Figure 6. Power dissipation analysis at different kink energies (a) Ek = 0.5 × 10-2 (meV) (b) Ek = 1.0 × 10-2 (meV) and (c) Ek = 1.5 × 10-2 (meV) Majority gate, there are some black spot cells available, as well as many less number of cells are there which dissipate power in black color. This shows the superiority of the proposed SRAM cell design compared to the previous designs in terms of power dissipation. Analysis of polarization for different input combinations is an interesting area to take care for a particular QCA based design because the proposed design must give the correct polarization for each of the input combinations. Figure 8 shows the polarization design for the proposed SRAM cell. The Blue color cell depicts logic ‘1’ while black color cell depicts logic ‘0’. Figure 8 shows the output polarization result for (1,1,1,1) input combination. From the figure, one can easily analyze that the proposed SRAM cell gives the correct polarization for a particular input combination. The same can be realized for other input combinations also. 05 10 15 20 25 30 35 40 45 Power in Kink Energy Types of power dissipation Power diss at Ek= 0.5 (meV) × 10-2 Previous design [14] Previous design [15] Proposed design 05 10 15 20 25 30 35 40 45 50 Power in Kink Energy Types of power dissipation Power diss at Ek = 1.0 (meV) × 10-2 Previous design [14] Previous design [15] Proposed design 0 10 20 30 40 50 60 Power in Kink Energy Types of power dissipation Power diss at Ek = 1.5 (meV) × 10-2 Previous design [14] Previous design [15] Proposed design 724 S. Kassa and S. Nema / IJE TRANSACTIONS B: Applications Vol. 32, No. 5, (May 2019) 720-725 Figure 7. Power dissipation analysis of proposed SRAM Cell Figure 8. Polarization analysis of the proposed SRAM cell for (1,1,1,1) input 5. CONCLUSION The proposed SRAM memory cell is designed by utilizing an open loop-based approach. It gives a proficient and robust end result SRAM memory cell structure as far as area, delay, cell counts and power consumption is compared with the reported designs. The proposed design structure saves the overall power up to 35.3% at maximum energy dissipation of circuit, 38.6% at average energy dissipation of circuit, 36.1% at minimum energy dissipation of circuit, 36.4% at average energy dissipation of circuit and 40.1% at average switching energy dissipation compared to the latest reported designs. Polarization analysis also supports the accuracy of the proposed SRAM cell design for different combination of inputs. The proposed design is energy efficient which can be considered as a base design for planning large systems similar to Microprocessors. 6. ACKNOWLEDGEMENT The authors duly acknowledge with gratitude the support of Dr. Sanjay Pawar, Principal, Usha Mittal Institute of Technology, SNDT Women’s University for his help and encouragement during the preparation of this manuscript. 7. REFERENCES 1. Yang, T., Kiehl, R. A., and Chua, L. O., “Tunneling phase logic cellular nonlinear networks”, International Journal of Bifurcation and Chaos, Vol. 11, No. 12, (2001), 2895–2911. 2. Likharev, K. K., “Single-electron devices and their applications”, Proceedings of the IEEE, Vol. 87, No. 4, (1999), 606–632. 3. Martel, R., Schmidt, T., Shea, H.R., Hertel, T., and Avouris, P., “Singleand multi-wall carbon nanotube field-effect transistors”, Applied Physics Letters, Vol. 73, No. 17, (1998), 2447–2449. 4. Lent, C. S., Tougaw, P. D., and Porod, W., “Quantum cellular automata: the physics of computing with arrays of quantum dot molecules”, In Proceedings Workshop on Physics and Computation, Dallas, USA, IEEE, (1994), 5–13. 5. Tougaw, P. D., “A device architecture for computing with quantum dots”, Proceedings of the IEEE, Vol. 85, No. 4, (1997), 541-557. 6. Kassa, S. R., and Nagaria, R. K., “A novel design of quantum dot cellular automata 5-input majority gate with some physical proofs”, Journal of Computational Electronics, Vol. 15, No. 1, (2016), 324–334. 7. Kassa, S. R. and Nagaria, R. K., “A Novel Design for 4-Bit Code Converters in Quantum Dot Cellular Automata”, Journal of Low Power Electronics, Vol. 13, No. 3, (2017), 482–489. 8. Zoka, S. and Gholami, M., “Two Novel D-Flip Flops with Level Triggered Reset in Quantum Dot Cellular Automata Technology”, International Journal of Engineering Transactions C: Aspects, Vol. 31, No. 3, (2017), 415–421. 9. Kassa, S