Reversible computing dissipates zero energy in terms of information loss at input and also it can detect error of circuit by keeping unique input-output mapping. In this paper, we have proposed a cost effective design of Reversible Programmable Logic Arrays (RPLAs) which is able to realize multi-output ESOP (Exclusive-OR Sum-Of-Product) functions by using a cost effective 3×3 reversible gate, called MG (MUX Gate). Also a new algorithm has been proposed for the calculation of critical path delay of reversible PLAs. The minimization processes consist of algorithms for ordering of output functions followed by the ordering of products. Five lower bounds on the numbers of gates, garbage and quantum costs of reversible PLAs are also proposed. Finally, we have compared the efficiency of proposed design with the existing one by providing benchmark functions analysis. The experimental results show that the proposed design outperforms the existing one in terms of numbers of gates, garbage, quantum costs and delay.
Design and minimization of reversible programmable logic arrays and its realization using pass transistors
1. Design and Minimization of Reversible
Programmable Logic Arrays and Its
Realization using Pass Transistors
Supervised by Dr. Hafiz Md. Hasan Babu
Professor, Dept. of Computer Science and Engineering
University of Dhaka, Dhaka-1000, Bangladesh
E-mails: hafizbabu@hotmail.com
Presented by Sajib Kumar Mitra
MS Student, Dept. of Computer Science and Engineering
University of Dhaka, Dhaka-1000, Bangladesh
E-mails: sajibmitra.csedu@yahoo.com
2. Purposes
• Define an new Architecture of RPLAs
• Minimization of Quantum Cost
• Reduction of Critical Path Delay
• Reduction of Number of Gates
• Garbage Outputs Optimization
• Pass Transistor Realization of RPLAs
3. Overview
• Reversible and Quantum Computing
• Reversible Programmable Logic Arrays
• Proposed Architecture of Reversible PLAs
• Delay Calculation of Reversible PLAs
• Pass Transistor Realization of MUX Gate
• Performance Analysis
• Conclusion
5. Reversible Computing
• Equal number of input states and output states
• Preserves an unique mapping between input and
output vectors for any Reversible circuit
• One or more operations can be united called
Reversible Gate
• (N x N) Reversible Gate has N number of inputs
and N number of outputs where N= {1, 2, 3, …}
[1] A. K. Biswas, M. M. Hasan, A. R. Chowdhury, and H. M. H. Babu, “Efficient approaches
for designing reversible binary coded decimalimplement in a single adders,”
Microelectronics Jounrnal, vol. 39, no. 12, pp. 1693–1703, December 2008.
6. • Limitation
• Feedback is strictly restricted
• Fan-out must be one always
Fig. 1: Basic difference between Irreversible and
Reversible Circuits
Reversible Computing…
8. In Quantum Computing, encode information as a series of
quantum-mechanical states such as spin directions of electrons
or polarization orientations of a photon that might represent
as or might represent a superposition of the two values.
Encoded data is represented by qubits rather than bits
which can perform certain calculations exponentially
faster than conventional computing.
10 βα +=q
Quantum Computing
[2]W. N. N. Hung, X. Song, G. Yang, J. Yang, and M. Perkowski, “Quantum logic synthesis by
symbolic reachability analysis,” in 41st
Conference on (DAC’04), Design Automation
Conference, May 2004, pp. 838–841.
9. Quantum Computation uses matrix multiplication rather than
conventional Boolean operations and the information
measurement is realized by calculation the state of qubits .
The matrix operations over qubits are simply specifies by using
quantum primitives. For example,
Fig. 3: Reversible behavior of Quantum matrix operation
Quantum Computing…
10. Input Output
A B P Q
0 0 0 0
0 1 0 1
1 0 1 1
1 1 1 0
Input/output
Pattern
Symbol
00 a
01 b
10 c
11 d
Quantum Computing…
Fig. 4: Working Principle of Unitary Controlled NOT (UCN)
12. Reversible Programmable Logic
Arrays
Reversible Programmable Logic Arrays was first proposed by A. R.
Chowdhury for multi-outputs function [4]. Reversible PLA
consists of following components:
1. Reversible AND Plane and 2. Reversible EX-OR Plane
Fig. 5: Reversible Programmable Logic Arrays
13. Reversible Programmable Logic
Arrays…
The existing design [4] of Reversible Programmable Logic Arrays is
shown in Fig. 6.
Fig. 6: Existing design of Reversible PLAs.
[4] Ahsan Raja Chowdhury, Rumana Nazmul, Hafiz Md. Hasan Babu: A New Approach to
Synthesize Multiple-Output Functions Using Reversible Programmable Logic Array.
VLSI Design 2006: 311-316
14. Reversible Programmable Logic
Arrays…
Limitation of Previous Design:
1. Toffoli gate produces huge number of unused
outputs which are same as primary inputs.
2. Used Conventional Architecture (Complement and
non-complement lines for copying input variables)
3. Requires huge number of Gates
16. Proposed Architecture of Reversible
PLAs
Proposed Design of RPLAs composed of followings:
1. EX-OR plane Optimization by using FG
2. Construction of AND plane by using MUX and FG
3. Delay Calculation based on Greedy Approach
We have used the following example to represent the
Proposed Design of RPLAs.
17. Proposed Architecture of Reversible
PLAs…
Construction of EX-OR plane :
1. EX-OR plane is constructed based on the ordering
of the size of Products of functions
2. EX-OR plane defines the particular order of all
products which will be followed by AND plane
Fig. 7: Optimized design of Reversible PLAs by using FG.
19. Proposed Architecture of Reversible
PLAs…
Sorted list of functions is: f4 f2 f1 f3 f5
f1 f2 f3 f4 f5
ab’ X X X
ab’c X X
a’b’c X
bc’ X X
ac X X X
Size of
Function
2 2 3 1 3
Functions
Products
20. Proposed Architecture of Reversible
PLAs…
f1 f2 f3 f4 f5
ab’ X X X
ab’c X X
a’b’c X
bc’ X X
ac X X X
Optimization of EX-OR Plane according to sorted list
of functions is:
f4 f2 f1 f3 f5
f4
ac
a’b’c
f2
4
f1
4ab’
ab’c
f3
4
4
bc’
4
4
2
f5
21. Proposed Architecture of Reversible
PLAs…
Theorem 1: Minimum number of Feynman gates to
realize EX-OR plane is: n + m – TDOT
Where ,
n= number of EX-OR
operations
m= number of output
functions
TDOT= total number of
cross-points
f4
ac
a’b’c
f2
4
f1
4ab’
ab’c
f3
4
4
bc’
4
4
2
f5
22. Proposed Architecture of Reversible
PLAs…
Construction of AND plane :
1.AND plane is constructed based on the ordering of the
Products
2. MUX and FG gates are used to design AND plane
3. Two different patterns of MUX gates have been used in
proposed design as follows:
Fig. 14: Reversible
MUX gate and two
different templates
of MUX gate which
are used in
proposed design.
26. Delay Calculation of Reversible PLAs…
We divide the calculation into two phases:
a. AND Plane Delay and
b. EX-OR Plane Delay
Then we have merged both of the delay respect to both
planes. In further realization of delay calculation, we
consider the following things:
a. Gate (Via) is represented as circle (DOT).
b. Delay of any gate is 1 and via (DOT) denotes 0.
c. Decimal value shows the delay of circle
28. Delay Calculation of AND Plane
Delay Calculation of Reversible PLAs…
Fig. 9: Delay Calculation of Reversible AND plane
29. Delay Calculation of AND Plane
Delay Calculation of Reversible PLAs…
Delay Calculation of Reversible AND plane
ac
a’b’c
5
ab’
ab’c
6
6
bc’
2
2
2
1 6
5
6
a b c
30. Delay Calculation of AND Plane
Delay Calculation of Reversible PLAs…
Delay Calculation of Reversible AND plane for each
Product line, APD (Pi)
APD (P4)= 2
1
a b c
Start 2
1 2 3
2 3
3 4 5
6
T
L
B
R
T
L
B
R
APD (P1)= 3
APD (P0)= 3
APD (P3)= 5
APD (P2)= 6
31. Delay Calculation of EX-OR Plane
Delay Calculation of Reversible PLAs…
Fig. 10: Delay Calculation of Reversible EX-OR plane
38. Conclusions
We have proposed a regular structure of Reversible
Programmable Logic Arrays (RPLAs) based on MUX and
Feynman logic and the focus of our design is as follows:
The garbage outputs as operational outputs that reduced
the number of AND operations in RPLAs.
AND plane based on the ordering of Products gives an
excellent throughput of the overall design.
The performance of the proposed design over the existing
one.
The experimental results show that the proposed design
outperforms the existing one in terms of numbers of gates,
garbages and quantum costs.