Binary adder with low transistor count

By introducing logic blocks and recovery blocks into binary adders, combined with improved AND gates and high-fan entry-AND gates, the shortcomings of existing adders in transistor count, area and power consumption are solved, and more efficient adders are achieved Designed to reduce latency and power consumption.

CN119032339APending Publication Date: 2024-11-26ISRAEL SUNILO CO LTD
View PDF 15 Cites 0 Cited by

Patent Information

Application Number
CN202380026031.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-03-08
Filing Date
2023-03-06
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Existing binary adders have shortcomings in terms of transistor count, semiconductor area and power consumption, especially the complexity of CMOS adders and large consumption, resulting in increased delay and area, and alternative topology or other types of logic schemes also face voltage degradation and The problem of degraded driving capability.

Method used

A binary adder with logic blocks and recovery blocks is designed, which is used to perform summing of binary inputs, and the recovery block is used to compensate for voltage level losses when the output is in a high or low logic state, through an improved AND gate and high fan-in-AND gate reduce the number of transistors and semiconductor area, reducing power consumption.

Benefits of technology

It realizes the reduction of transistor count and semiconductor area, reduces power consumption and delay, and is suitable for designing multi-bit adders, improves signal integrity and driving power, and avoids the problems of voltage degradation and area increase.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119032339A_ABST
    Figure CN119032339A_ABST
Patent Text Reader

Abstract

A method for implementing a logic circuit employing a combination of binary adders and AND gates of different lengths with summing and carry outputs, comprising the steps of replacing at least one known CMOS implemented binary adder with an improved binary adder consisting of: a logic block comprising a plurality of logic blocks, the summation module is used for performing summation between binary inputs of the logic blocks; and a recovery block connected between an output of the logic block and an output of the binary adder circuit for compensating for a voltage level loss when the output is in a high logic state or a low logic state. At least one CMOS AND gate is replaced with a high fan-in AND gate above a predetermined threshold by an improved AND gate, the improved AND gate consisting of: a logic block for performing an AND operation between binary inputs of the logic block; and a recovery block connected between the output of the logic block and the output of the AND gate for compensating the voltage level loss when the output is in the high logic state or the low logic state.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of static logic digital circuits. More particularly, the present invention relates to an efficient design of binary adders which allows reducing the number of required transistors, semiconductor area and power consumption. Background Art

[0002] Binary adders are the main components used in VLSI circuits for arithmetic operations, multiplexing, and digital filters. The circuit performance of a multi-bit adder depends on the design of the basic adder, as well as the topology of multiple basic adders that form the multi-bit adder; i.e., cascaded stages, tree structure, etc.

[0003] CMOS gates outperform single-type MOSFETs in terms of power consumption, but are inferior in terms of packaging density and speed. CMOS gates are also limited to relatively small fan-in (the number of inputs that a gate can handle - in most cases a maximum of four inputs), so functions more advanced than simple logic (e.g., arithmetic operations) require cascading multiple levels of CMOS gates, which increases semiconductor area as well as power consumption and latency. In addition, advanced CMOS technology nodes (i.e., from 65nm down to 3nm) suffer from the problem of high static power consumption (due to subthreshold and junction leakage). Because CMOS adders are complex and consume relatively large semiconductor areas, other alternatives have been sought via alternative topologies or other types of logic.

[0004] Mehedi et al. conducted a comprehensive study of 1-Bit full adder cells: review, performance comparison, and scalability analysis (Comprehensive study of 1-Bit full adder cells: review, performance comparison, and scalability analysis, Mehedi Hasan, Abdul Hasib Siddique, Abdal Hoque Mondol, Mainul Hossain, Hasan U. Zaman & Sharnali Islam. SN Applied Sciences volume 3, Article number: 644, 2021), in which they discussed the advantages and disadvantages of single-transistor logic adders, such as complementary channel logic (CPL) where voltage swing degradation is a key issue, or transmission gate (TG) based adders where the problem of voltage swing degradation has been solved but drive power difference is the main problem. Such TG adders are described in US4559609. Hybrid logic based full adder circuits that attempt to solve voltage swing degradation suffer from increased area, higher transistor count, or both. Recent gate diffusion input (GDI) adders, such as the adder described by Arkadiy Morgenshtein et al. (Arkadiy Morgenshtein et al., Full-Swing Gate Diffusion Input logic-Case-study of low-power CLA adder design, Integration-the VLSI journal Vol 47, pp.62-70, 2014) also suffer from voltage degradation and reduced drive capability. Mehedi et al. analyzed other implementations of the adder, which all suffer from one or more of the above-mentioned shortcomings. In general, adders based on channel transistors (such as the adder described in US 4905180) are at a disadvantage compared to CMOS adders because NMOS transistors are weak conductors of logic "1" and PMOS transistors are weak conductors of logic "0". Transmitting logic "1" through NMOS transistors causes impaired signal integrity due to the voltage drop at the output (Vdd-Vtn instead of Vdd). Similarly, transmitting a logic "0" through a PMOS transistor results in |Vtp| at the output instead of zero voltage (ie, ground).

[0005] The most basic CMOS one-and-a-half-bit adder (1bHA) consists of a two-input XOR gate for computing the sum and a two-input AND gate for computing the carry-out of the addition of two 1-bit numbers. For example, the US4054788 patent describes one of many implementations of the 1bHA that improves latency at the expense of increased transistor count (relative to CMOS) and increased complexity (i.e., resistors). Designing a 1bHA circuit that outperforms CMOS in all performance metrics (i.e., area, power, and latency) is challenging.

[0006] The most compact CMOS one-bit full adder (1bFA) contains 28 transistors. As discussed by Mehedi et al., alternatives that include fewer transistors are inferior in terms of signal integrity, drive power, semiconductor area, or a combination thereof.

[0007] In addition to the one-bit adder, the four-bit adder is most commonly used in the design of multi-bit adders (i.e., 8-bit, 16-bit, 32-bit, and 64-bit). The four-bit adder circuit either exhibits a regular repetitive structure by linking or cascading one-bit adders, or gains increased area and complexity (irregular structure) for increasing speed. There are a large number of four-bit adders with different architectures in the literature. US4439835 and US 6978290B2 describe a carry ripple adder (CRA) by connecting four 1bFAs in series. It is the most basic four-bit adder and has a simple and regular structure. However, CRA is slower for multi-bit addition (e.g., 16-bit) of binary numbers. It is characterized by O (n) area and delay (where n is the number of bits of addendum). US3100835, US5027312, and US5396445 describe a carry select adder (CSelA) that improves the delay of CRA. It consists of two parallel ripple-carry adders and a multiplexer. CSelA has an area of ​​O(n) and latency, and the same is true for the carry skip adder described in US6199091B1 and the carry increment adder described in US5548546. US7111033B2 proposes a carry save adder (CSA) featuring an O(n) area and improved O(log(n)) latency. US5964827 describes a fast carry lookahead adder (CLA) featuring an increased area of ​​O(log(n)) in exchange for lower O(log(n)) latency. However, it suffers from the problem of irregular structure, which complicates its layout due to wiring congestion. It is worth noting that if there is no restriction on fan-in / fan-out, then the CLA adder can be implemented with only two gate levels. The CLA belongs to the family of parallel prefix adders (PPAs), to which the Kogge-Stone adder (described in Kogge, Peter Michael and Stone, Harold SA Parallel Algorithm for the Efficient Solution of a General Class of Recurrence Equations. IEEE Transactions on Computers. C-22(8):786-793, 1973) and the Brent-Kung adder (described in Brent, Richard Peirce and Kung, Hsiang Te. A Regular Layout for Parallel Adders. IEEE Transactions on Computers. C-31(3):260-264, 1982) belong, and which are widely used despite their increased area and layout defects. US5504915 describes a Wallace tree adder (WTA) of multi-stage adders. However, tree structured adders are relatively slow and less common, and there are more exotic adders such as neural network adders (described in US5016211).

[0008] It is therefore an object of the present invention to provide a one-bit and four-bit adder circuit implementation which reduces the number of transistors required compared to similar implementations using known CMOS technology.

[0009] It is another object of the present invention to provide a one-bit and four-bit adder circuit implementation that reduces power consumption.

[0010] Another object of the present invention is to provide any multi-bit adder implementation that saves semiconductor area, reduces latency and power consumption.

[0011] Other objects and advantages of the invention will become apparent as the description proceeds. Summary of the invention

[0012] A 1-bit binary adder with sum and carry outputs having reduced transistor count and semiconductor area, comprising:

[0013] a) a logic block for performing a summation between binary inputs of the logic block; and

[0014] b) A recovery block connected between the output of the logic block and the output of the binary adder circuit to compensate for the voltage level loss when the output is in a high logic state or a low logic state.

[0015] A multi-bit adder is implemented by combining the 1-bit binary adders described above.

[0016] The voltage at the sum and carry outputs may be compatible with either a "1" logic level or a "0" logic level.

[0017] In one aspect, the number of transistors required to implement a binary adder is less than 28.

[0018] The binary adder according to claim 1 can be implemented as a full adder (FA) or a half adder (HA) or a combination thereof.

[0019] A recovery block can consist of:

[0020] - Standard CMOS inverter;

[0021] - Standard CMOS buffer;

[0022] - Schmitt trigger;

[0023] - combinations thereof.

[0024] A method for implementing a logic circuit using a combination of binary adders of different lengths and AND gates with sum and carry outputs, comprising the steps of:

[0025] a) replacing at least one known CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of:

[0026] a.1) a logic block for performing a summation between binary inputs of the logic block;

[0027] a.2) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for the voltage level loss when the output is in a high logic state or a low logic state;

[0028] b) replacing at least one CMOS AND gate with a high fan-in AND gate above a predetermined threshold by an improved AND gate, the improved AND gate consisting of:

[0029] b.1) a logic block for performing an AND operation between binary inputs of the logic block; and

[0030] b.2) A recovery block, connected between the output of the logic block and the output of the AND gate, for compensating for the voltage level loss when the output is in a high logic state or a low logic state.

[0031] A method for implementing a logic circuit using a combination of binary adders of different lengths and AND gates with sum and carry outputs, comprising the steps of:

[0032] a) replacing at least one known CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of:

[0033] a.1) a logic block for performing a summation between binary inputs of the logic block;

[0034] a.2) The recovery block is connected between the output of the logic block and the output of the binary adder circuit to compensate for the voltage drop when the output is in a high logic state or a low logic state.

[0035] Flat loss;

[0036] a.3) Using improved AND gates to replace the AND gates with high fan-in AND gates above a predetermined threshold

[0037] With one less CMOS AND gate, the improved AND gate consists of the following:

[0038] b.1) a logic block for performing an AND operation between binary inputs of the logic block; and

[0039] b.2) A recovery block, connected between the output of the logic block and the output of the AND gate, for compensating for the voltage level loss when the output is in a high logic state or a low logic state.

[0040] The high fan-in AND gate also includes a pull-down block connected between the logic block and the output of the AND gate for further discharging a voltage corresponding to a high logic state to ground in addition to discharging via an inherent current leakage path after a logic operation entailing a low logic state.

[0041] The drop-down block can be selected from the following group:

[0042] -diode;

[0043] - a transistor configured to operate as a diode;

[0044] - a plurality of transistors configured to operate as diodes;

[0045] - A combination of PMOS and NMOS transistors that acts as a diode.

[0046] The logic block of the high fan-in AND gate may be a stack of transistors implementing the connection of the AND gate, or a parallel connection of transistors implementing the AND gate, or a combination thereof.

[0047] The predetermined fan-in threshold may be three.

[0048] A multi-bit adder is implemented by combining the 1-bit binary adder described above and at least one high fan-in AND gate.

[0049] A multi-bit binary adder with a sum and carry output having a reduced transistor count and semiconductor area, comprising a combination of two or more 1-bit and / or 4-bit binary adders connected in parallel or in series, wherein each 1-bit and / or 4-bit adder comprises:

[0050] a) a logic block for performing a summation between binary inputs of said logic block; and

[0051] b) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for a voltage level loss when the output is in a high logic state or a low logic state.

[0052] The numbers added can contain any number of bits and can be connected in a hierarchical architecture.

[0053] A binary adder may be used to perform a subtraction operation, for example, using 2's complement.

[0054] Unless otherwise defined, all technical and / or scientific terms used herein have the same meanings as those of ordinary skill in the art to which the invention belongs. Although methods and materials similar or equivalent to the methods and materials described herein may be used in practice or testing embodiments of the present invention, exemplary methods and / or materials are described below. In the event of a conflict, the patent specification (including definitions) shall prevail. In addition, materials, methods and examples are merely illustrative and are not intended to necessarily limit. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The above and other features and advantages of the present invention will be better understood through the following illustrative and non-limiting detailed description of preferred embodiments of the present invention in conjunction with the accompanying drawings, in which:

[0056] Figure 1 is a circuit diagram of a one-and-a-half-bit adder using a single transistor as a logic element according to the present invention. According to an embodiment of the present invention, the one-and-a-half-bit adder is composed of 12 transistors;

[0057] Figure 2 is another implementation of a one-and-a-half-bit adder combining single-transistor logic and known CMOS logic consisting of 12 transistors;

[0058] Figure 3 is a circuit diagram of the most compact known CMOS low voltage one-bit full adder, which includes 14 CMOS pairs, i.e., 28 transistors;

[0059] Figure 4 is a circuit diagram of a hybrid one-bit full adder of the present invention, the one-bit full adder comprising 18 transistors combined with CMOS logic;

[0060] Figure 5 is a circuit diagram of a one-bit full adder of the present invention, the one-bit full adder comprising 22 transistors;

[0061] Figure 6 is another circuit diagram implementation of a one-bit full adder of the present invention, the one-bit full adder comprising 16 transistors;

[0062] Figure 7 is a block diagram of a three-input AND gate as described in International Patent Application No. PCT / IL2022 / 050981;

[0063] Figure 8 (Prior Art) is a block diagram of a four-bit carry ripple adder (CRA) architecture, comprising four one-bit full adders connected in series;

[0064] Fig. 9 (Prior Art) is a block diagram of a 4-bit carry-save adder (CSA) architecture;

[0065] Fig.10 (Prior Art) is a circuit diagram of the basic architecture of a parallel prefix adder known as a carry-lookahead adder (CLA); and

[0066] Fig.11 is a block diagram illustrating an embodiment of a CLA having a half adder and a full adder of the present invention. DETAILED DESCRIPTION

[0067] The present invention describes low transistor count static 1-bit and 4-bit basic adders that can be used to design any multi-bit adder to achieve area savings, reduced power consumption, and reduced latency. The 1-bit adder circuit and the 4-bit adder circuit provided by the present invention use fewer transistors than the same circuit implemented using existing CMOS technology.

[0068] The present invention relates to digital circuits, and in particular to binary adders using single-transistor based logic, i.e., circuits in which the addition function is implemented by non-CMOS logic. The method of the present invention can be implemented for any adder architecture, including but not limited to parallel prefix adders (PPA - adders that use prefix operations for efficient addition), such as Kogge-Stone adders, Brent-Kung adders (which are parallel prefixes in the form of CSA adders), tree adders (such as Wallace tree adders), etc.

[0069] Multiple adders can also be combined in series or in parallel to form addition, subtraction, multiplication, and division circuits. Figure 1-Figure 11 , wherein like numerals are used for like and corresponding parts, for a better understanding of the preferred embodiments of the present invention and its advantages.

[0070] A half adder with binary inputs A and B computes the sum output as follows:

[0071]

[0072] The bar in equation (1) represents complementary values ​​and the plus sign represents a logical OR. Equation (1) is equivalent to an exclusive OR (XOR) between A and B, represented by the symbol ⊕.

[0073] The carry output of the half adder is:

[0074] C out =AB (Equation 2)

[0075] Equation (2) is equivalent to the logical AND operation of binary numbers A and B.

[0076] Referring now to the drawings, and in particular to the Figure 1 , Figure 1 A one-and-a-half-bit adder using single-transistor logic.

[0077] A one-and-a-half adder combines two one-bit binary numbers to form their sum and carry, wherein any carry from the previous operation is ignored. Two one-bit inputs 1 and 2 are connected to multiple gates of a single-type transistor that computes binary addition. A drive voltage 3 provides a supply voltage to the adder. According to the present invention, a novel one-and-a-half adder implementation is provided, comprising a first block and a second block. The first block is composed of a plurality of single-type transistors 4-11 to perform the addition of two one-bit binary numbers (inputs 1, 2). The second block includes transistors 12-15 and is adapted to recover the signals generated by the first block and output the sum (Sum) and carry output (Cout) to acceptable voltage levels. In an embodiment of the half adder, the threshold voltage and channel width of transistors 4-5 may be different from the threshold voltage and channel width of transistors 6-7 or 8-9 and transistors 10-11. In another embodiment, the threshold voltage and channel width of transistors 12, 13 or 14, 15 may be different from the threshold voltage and channel width of transistors 4-11. In Figure 1 In another embodiment of the adder depicted in , transistors 4, 5 have a higher threshold voltage that is different from the threshold voltage of transistors 6, 7. In another embodiment, the threshold voltage and channel width of transistors 8 and 9 are different from the threshold voltage and channel width of transistors 10-11, and similarly for transistors 4, 5 and transistors 6, 7. Using multiple threshold voltages and multiple channel widths allows the half adder to be tuned to output acceptable voltages and reduce power consumption.

[0078] In an embodiment, a one-and-a-half adder is implemented by connecting two one-bit inputs to multiple gates of a single type of transistor that calculates binary addition. Then, the sum is generated by a first block of the one-and-a-half adder (composed of a first set of transistors). The carry output is generated by a second block of the one-and-a-half adder (composed of a second set of transistors).

[0079] Figure 2 Another embodiment of a half adder according to the present invention is given. Two one-bit inputs 18 and 19 are connected to a plurality of single-type transistors that calculate binary addition. A drive voltage Vdd provides a supply voltage for the adder. According to the present invention, a novel one-bit half adder embodiment is provided, comprising a first block composed of transistors 20-25 for generating a sum 32 and a second block composed of transistors 26-31 for generating a carry output 33.

[0080] Figure 3The most compact low voltage 1-bit full adder of known CMOS is implemented, which includes 28 transistors. The number of transistors required to implement a binary adder can be less than 28. Two 1-bit numbers 34, 36 and a carry-in 35 are fed into the circuit to produce the sum of the addends 37 and the carry-out 38 (A+B|carry-in=sum|carry-out). Figure 3 In the embodiment of the present invention, a one-bit full adder is formed by cascading two one-bit half adders. The driving voltage Vdd provides a supply voltage for the adder.

[0081] The single bit binary numbers (A, B) 34 and 36 and the carry in (Cin) 35 from the previous stage are added to form a sum 37 and a carry out 38 as given below:

[0082] Sum=(A⊕B)⊕C in (Equation 3)

[0083] C out =AB+C in (A⊕B) (Equation 4)

[0084] Different Boolean representations of equations (3)-(4) are given below:

[0085] Sum=(A⊙B)⊙C in (Equation 5)

[0086]

[0087] The symbol ⊙ represents the exclusive-OR (XNOR) operator.

[0088] Figure 4 A one-bit full adder implementation including 18 transistors based on equations (5)-(6) and including a first block and a second block using single-transistor logic according to an embodiment of the present invention is illustrated. Two one-bit binary numbers 39 and 40 are added, including a carry input 41 from a previous stage. The first block is composed of a plurality of single-type transistors 42-51 to calculate the carry output of the addition of the two one-bit binary numbers. Transistors 42-45 perform an XNOR operation on inputs 39 and 40 (i.e., A⊙B), and their outputs are fed to transistors 46-49 for continuous XNOR operations with input 41 (carry input) to obtain the sum 60 of inputs 39 and 40 according to equation (5).

[0089] Transistors 50-51 act as selectors and output carry outputs based on inputs 39-41 (A, B, carry in). Whenever inputs 39, 40 are ("0", "0") or ("1", "1"), selectors 50, 51 output input 39 as carry out; that is, Cout = A and perform Boolean operation A (A⊙B).

[0090] Whenever input 39, 40 has odd "1" parity; that is, either ("1", "0") and ("0", "1"), the output of selector 50, 51 is the input carry; that is, the Boolean operation is performed And Cout=Cin.

[0091] The sum and output carry are provided at the output of the second block (recovery block), which is composed of transistors 52-59 to compensate for the voltage level loss when the output is in a high logic state or a low logic state. The recovery block can be composed of a standard CMOS inverter, a standard CMOS buffer, a Schmitt trigger, or any combination thereof.

[0092] The second block is arranged to recover the sum and carry output signals generated by the first block and output the sum and carry outputs at acceptable voltage levels. In an embodiment of a full adder, the threshold voltage and channel width of transistors 42-49 may be different from the threshold voltage and channel width of transistors 50, 51 (implementing the selector) and transistors 52-59. Figure 4 In another embodiment of the adder depicted in , transistors 42, 43 or transistors 46, 47 have a threshold voltage different from the threshold voltage of transistors 44, 45 or 48, 49 or 50, 51. Using multiple threshold voltages and multiple channel widths is suitable for tuning the full adder to output an acceptable voltage and reduce power consumption. The drive voltage Vdd provides a supply voltage for the adder.

[0093] Figure 5 6 is another embodiment of a one-bit full adder according to the present invention, which includes 22 transistors based on equations (3)-(4) and includes a first block and a second block using single-transistor logic. Two one-bit binary numbers A (62) and B (63) are added, including a carry input 64 from the previous stage. The first block is composed of multiple single-type transistors to perform a carry output for the addition of two one-bit binary numbers. Transistors 65-70 perform an XOR operation on the input (i.e., A⊕B), and their outputs are fed to transistors 71-76 to perform a continuous XOR operation with the carry input of the input to obtain the sum 83 of inputs 62-64 according to equation (3).

[0094] Transistors 75-82 act as AND-OR gates that determine the carry-out result based on equation (4). Transistors 75-78 perform a logical AND function between inputs A 62 and B 63, and transistors 79-82 perform a logical AND function between carry-in 64 and A⊕B. The combined output of these two AND gates produces an AND-OR circuit that generates a carry-out.

[0095] Figure 6 is another embodiment of a one-bit full adder according to the present invention, which includes 16 transistors using transmission gates 90, 91 and 94, 95 and 96-99, and is based on equations (3)-(4). Two one-bit binary numbers A 85 and B 86 are added, including the carry input 87 from the previous stage. Transistors 88-91 perform an XOR operation between inputs A 85 and B 86 (i.e., A⊕B), and their outputs are fed to transistors 92-95 for continuous XOR operations with the input (carry input) to obtain the sum of the inputs 100 according to equation (3).

[0096] Pass gate transistors 96-99 act as AND-OR gates that determine the carry output result based on equation (4). Pass gate transistors 96-97 generate a logical AND function between inputs A and B, and pass gate transistors 98-99 perform a logical AND function between Cin and A⊕B. The combined output of these two AND gates produces an AND-OR circuit that generates a carry output 101.

[0097] Figure 7 is a block diagram of a three-input AND gate as described in International Patent Application No. PCT / IL2022 / 050981 (of the same applicant). Figure 7 Embodiments of the multi-input AND gate are used in some of the multi-bit full adders described in this invention. Figure 7 In one embodiment of the circuit, inputs 102-104 are fed into the gate of a stack of n-type transistors operating as a logic AND gate. When all transistors are in the "ON" state, the supply voltage Vdd is transmitted to the wire 105. The recovery block 106b compensates for the threshold voltage drop of the stack 102-104. When any of the inputs 102-104 is "0", the pull-down block 106a pulls the voltage on the wire 105 to ground.

[0098] Figure 8is a general block diagram of a four-bit carry ripple adder (CRA architecture) as described in USP 4439835 and USP 6978290B2, which includes four one-bit full adders 122-125 in series. The carry input 108 of the first adder is "0" (i.e., ground), but can also take "1". The carry input of any other adder in the series is fed with the carry output of the previous adder. Two four-bit numbers are added. The first addend bits are 109, 111, 113, 115 (109 is LSB and 115 is MSB), and the second addend bits are 110 (LSB), 112, 114, 116 (MSB). The summand bits are 117 (LSB), 118, 119, 120 (MSB), and the carry output is 121. Prior art elements 122-125 include Figure 3 In an embodiment of the present invention, Figure 3 The known CMOS one-bit full adder and Figure 4-Figure 6 A single or multiple mix and match placement of one-bit full adders results in a CRA topology with fewer transistors, reduced latency, and reduced power consumption. In another embodiment of the present invention, comprising Figure 4-Figure 6 The CRA of all elements 122-125 of the full adder results in fewer transistors, reduced latency, and reduced power consumption.

[0099] Fig. 9 (Prior Art) A four-bit carry-save adder (CSA) as described in US7111033B2 is given. Two four-bit numbers are added. The first addend bits are 134-137 (134 is the LSB, 137 is the MSB), and the second addend bits are 138 (LSB), 139, 140, 141 (MSB). The sum terms are 142 (LSB), 143, 144, 145 (MSB), and the carry out is 146. Elements 126-130 are formed by a known CMOS one-and-a-half-bit adder (i.e., Figure 3 The full adder of FIG. 1 is composed of a full adder of FIG. 1 , wherein the carry input is grounded), and elements 131-133 are composed of known Figure 3 In the embodiment of the present invention, a single or multiple CMOS one-bit full adder is respectively Figure 1-Figure 2 and Figure 3-Figure 6 The mixing and matching placement of the half adders and full adders of the present invention in the CSA results in fewer transistors, reduced latency and reduced power consumption. In another embodiment of the present invention, comprising Figure 1-Figure 2 The half adder and Figure 4-Figure 6 The CSA of all elements 126-130 of the full adder results in a CSA topology with fewer transistors, reduced latency, and reduced power consumption.

[0100] Fig.10 A basic parallel prefix adder of a four-bit carry-lookahead adder (CLA) as described in US5964827 is depicted in FIG. The first addend bits are 147-150 (147 is LSB, 150 is MSB), and the second addend bits are 151 (LSB), 152, 153, 154 (MSB). The carry-in bit is C0 (155). The sum term bits are 156 (LSB), 157, 158, 159 (MSB), and the carry-out is 160. Gates 161-163 are high fan-in CMOS AND gates. In an embodiment of the invention, Figure 7 A mix and match of single or multiple AND gates is introduced as an alternative to the known CMOS AND gates 161-163 to reduce the number of transistors, thereby saving area, shortening the critical path of the adder and reducing power consumption. As given in the international application PCT / IL2022 / 050981 (of the same applicant), five, four or three input AND gates can be constructed. In another embodiment of the present invention, all AND gates 161-163 are replaced by Figure 7 The equivalent gate of the AND gate can further save area, shorten the critical path of the adder, and reduce power consumption.

[0101] Fig.11 Given Fig.10 Different block diagrams of the same four-bit CLA using half adders and full adders implemented according to the present invention. The carry input bit of the input is 182. Two four-bit numbers A and B (A inputs 164-167 and B inputs 168-171) are added. The sum terms are 183 (LSB), 184, 185, 186 (MSB), and the carry output is 187. Gates 175-177 are high fan-in CMOS AND gates of 3 or more inputs. Prior art elements 172-174 are known CMOS one-bit half adders, and 178-181 are known CMOS one-bit full adders.

[0102] In an embodiment of the present invention, one or more Figure 1-Figure 2 The half adder of the present invention, Figure 3-Figure 6 The full adder and Figure 7 The mixing and matching of equivalent AND gates with known CMOS circuits can reduce transistor counts; thereby saving area, shortening the critical path of the adder, and reducing power consumption. In another embodiment of the present invention, all half adders 172-174, all high fan-in AND gates 175-177, and all full adders 178-181 are Figure 1-Figure 2 The half adder of the present invention, Figure 4-Figure 6 The full adder and Figure 7 The equivalent AND gate is used instead.

[0103] According to another embodiment, addition using a higher number of bits (such as 8-bit addition, 16-bit addition, 32-bit addition, etc.) is performed by using any combination of 1-bit adders and / or 4-bit adders to form adders with a higher number of bits and adders connected in a hierarchical architecture (for reducing delay). In another embodiment, Figure 8-Figure 11 The connection of two, four or eight 4-bit adders can produce an 8-bit adder, a 16-bit adder and a 32-bit adder, respectively. In another embodiment, Figure 1-Figure 6 The appropriate connection of the adders can form any adder of any size. In another embodiment, Figure 1-Figure 6 The adder and Figure 8-Figure 11 The connection of adders can form any adder of any size. In another embodiment, the adder of the present invention can be used in any multi-bit adder architecture.

[0104] In one embodiment, a multi-bit binary adder has a sum and carry output and comprises a combination of two or more 1-bit and / or 4-bit binary adders connected in parallel or in series. Each 1-bit and / or 4-bit adder comprises a logic block for performing a sum between binary inputs of the logic block; and a recovery block connected between the output of the logic block and the output of the binary adder circuit for compensating for voltage level loss when the output is in a high logic state or a low logic state.

[0105] Although the above embodiments relate to addition operations, as will be appreciated by those skilled in the art, the present invention may also be used for subtraction operations. For example, subtraction of two binary numbers may be accomplished by adding a 2's complement (a 2's complement is a mathematical operation used to reversibly convert a positive binary number into a negative binary number of equivalent value, using the binary digit with the largest position value to indicate whether the binary number is positive or negative. In the first step, a given decimal number is converted into binary. In the next step, the one's complement of the binary number is obtained by converting each "0" into a "1" and converting a "1" into a "0". In the next step, a "1" is added to the one's complement. When the most significant bit is "1", the number is marked as the negative of the subtrahend to the minuend, and the final carry (if any) is ignored. If the MSB bit in the result of the addition is "0", then the result of the addition is the correct answer. If the MSB bit is "1", then this means that the answer has a negative sign. In this case, the true magnitude is given by the 2's complement of the result of the addition.

[0106] In one embodiment, a logic circuit using a combination of a binary adder and an AND gate with a sum and carry output is implemented in the following manner: replacing at least one CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of: a logic block for performing a sum between binary inputs of the logic block; a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for a voltage level loss when the output is in a high logic state or a low logic state; replacing at least one CMOS AND gate with a high fan-in AND gate above a predetermined threshold through an improved AND gate, the improved AND gate consisting of: a logic block for performing an AND operation between binary inputs of the logic block; a recovery block connected between the output of the logic block and the output of the AND gate, for compensating for a voltage level loss when the output is in a high logic state or a low logic state.

[0107] In one embodiment, a logic circuit using a combination of a binary adder and an AND gate with a sum and carry output is implemented in the following manner: replacing at least one CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of: a logic block for performing a sum between binary inputs of the logic block; a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for a voltage level loss when the output is in a high logic state or a low logic state; replacing at least one CMOS AND gate with a high fan-in AND gate above a predetermined threshold using an improved AND gate, the improved AND gate consisting of: a logic block for performing an AND operation between binary inputs of the logic block; and a recovery block connected between the output of the logic block and the output of the AND gate, for compensating for a voltage level loss when the output is in a high logic state or a low logic state.

[0108] It should be appreciated that certain features of the invention described in the context of separate embodiments for the sake of clarity may also be provided in combination in a single embodiment. Conversely, various features of the invention described in the context of a single embodiment for the sake of brevity may also be provided separately or in any suitable sub-combination or in any other described embodiment of the invention as appropriate. Certain features described in the context of individual embodiments are not considered essential features of those embodiments unless the embodiment is inoperable without those elements.

[0109] Although the present invention has been described in conjunction with the specific embodiments of the present invention, it is obvious that many substitutions, modifications and variations will be obvious to those skilled in the art. Therefore, it is intended to cover all such substitutions, modifications and variations that fall within the spirit and broad scope of the appended claims.

[0110] All publications, patents and patent applications mentioned in this specification are hereby incorporated by reference in their entirety into this specification, to the same extent as if each individual publication, patent or patent application was expressly and individually indicated as being incorporated by reference herein. In addition, the citation or identification of any reference in this application should not be construed as an admission that such reference can be used as prior art to the present invention. To the extent that section headings are used, they should not be construed as necessarily limiting.

Claims

1. A 1-bit binary adder with a sum and carry output, the 1-bit binary adder having a reduced transistor count and semiconductor area, the 1-bit binary adder include: a) a logic block for performing a summation between binary inputs of said logic block; as well as b) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for a voltage level loss when the output is in a high logic state or a low logic state.

2. A multi-bit adder, wherein the multi-bit adder is implemented by combining several 1-bit binary adders according to claim 1.

3. The binary adder of claim 1 , wherein the voltage at the sum and carry outputs is compatible with a “1” logic level or a “0” logic level.

4. The binary adder of claim 1, wherein the number of transistors required to implement the binary adder is less than 28.

5. The binary adder according to claim 1, wherein the number of transistors required to implement the binary adder is 18.

6. The binary adder of claim 1, implemented as a full adder (FA).

7. The binary adder of claim 1, implemented as a half adder (HA).

8. The binary adder according to claim 1, wherein the recovery block is composed of the following composition: -Standard CMOS inverter; - Standard CMOS buffer; - Schmitt trigger; - combinations thereof.

9. A multi-bit adder, the multi-bit adder being implemented by a combination of the 1-bit binary adder according to claim 1 and at least one high fan-in AND gate.

10. The binary adder according to claim 1, wherein the one-and-a-half-bit adder is implemented by: a) Connect two one-bit inputs to multiple gates of a single type of transistor that compute binary addition; b) generating a sum by a first block of said one-bit half adders composed of a first set of transistors; c) generating a carry output by a second block of said one-bit half adder composed of a second set of transistors.

11. The binary adder of claim 1 , wherein the one-bit full adder comprises a first block and a second block using single transistor logic, in: -Add two single-bit binary numbers, including the carry input from the previous stage; - the first block is composed of a plurality of single-type transistors to perform a carry output for the addition of two one-bit binary numbers; - Perform an XOR operation on the input; - Perform consecutive XOR operations with the carry-in input to obtain the sum of the inputs.

12. A multi-bit binary adder with sum and carry outputs, the multi-bit binary adder having a reduced transistor count and semiconductor area, the multi-bit binary adder comprising a combination of two or more 1-bit and / or 4-bit binary adders connected in parallel or in series, wherein each 1-bit and / or 4-bit adder include: a) a logic block for performing a summation between binary inputs of said logic block; as well as b) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for a voltage level loss when the output is in a high logic state or a low logic state.

13. The multi-bit binary adder of claim 13, wherein the numbers added contain any number of bits.

14. The multi-bit binary adder of claim 13, wherein the 1-bit binary adders are connected in a layered architecture.

15. The 1-bit binary adder according to claim 1 or 13, wherein the 1-bit binary adder is used to perform a subtraction operation.

16. The 1-bit binary adder of claim 16, wherein the subtraction operation is performed using 2's complement.

17. A method for implementing a logic circuit, the logic circuit using a combination of a binary adder with a sum and carry output and an AND gate, the method include: a) replacing at least one CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of: a.1) a logic block for performing a summation between binary inputs of said logic block; a.2) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for voltage level loss when the output is in a high logic state or a low logic state; b) replacing at least one CMOS AND gate with a high fan-in AND gate above a predetermined threshold by an improved AND gate, the improved AND gate consisting of: b.1) a logic block for performing an AND operation between binary inputs of said logic block; and b.2) A recovery block connected between the output of the logic block and the output of the AND gate, for compensating for voltage level loss when the output is in a high logic state or a low logic state.

18. A method for implementing a logic circuit, the logic circuit using a combination of a binary adder with a sum and carry output and an AND gate, the method include: a) replacing at least one CMOS implemented binary adder with an improved binary adder, the improved binary adder consisting of: a.1) a logic block for performing a summation between binary inputs of said logic block; a.2) a recovery block connected between the output of the logic block and the output of the binary adder circuit, for compensating for voltage level loss when the output is in a high logic state or a low logic state; a.3) replacing at least one CMOS AND gate with a high fan-in AND gate above a predetermined threshold using an improved AND gate, the improved AND gate consisting of: b.1) a logic block for performing an AND operation between binary inputs of said logic block; and b.2) A recovery block connected between the output of the logic block and the output of the AND gate, for compensating for voltage level loss when the output is in a high logic state or a low logic state.

19. The method of claim 18, wherein the AND gate further comprises a pull-down block connected between the logic block and the output of the AND gate, for further discharging a voltage corresponding to a high logic state to ground in addition to discharging via an inherent current leakage path after a logic operation causing a low logic state.

20. The method of claim 19, wherein the drop-down block is selected from the group consisting of: -diode; - a transistor configured to operate as a diode; - a plurality of transistors configured to operate as diodes; - A combination of PMOS and NMOS transistors that acts as a diode.

21. The method of claim 18, wherein the logic block of the AND gate is a stack of transistors implementing the connection of the AND gate, or a parallel connection of transistors implementing the AND gate, or a combination thereof.

22. The method of claims 18-20, wherein the predetermined fan-in threshold is three.

23. The method of claim 18 or 19, wherein the binary adder is a full adder (FA) or a half adder (HA) or a combination thereof.

Citation Information

Patent Citations

  • Selecting adder

    US3100835A

  • Modular binary half-adder

    US4054788A

  • Apparatus for and method of generation of ripple carry signals in conjunction with logical adding circuitry

    US4439835A

  • Full adder using transmission gates

    US4559609A

  • MOS adder with minimum pass gates in carry line

    US4905180A