Chips and computing systems

By introducing fault-tolerant logic circuits and intentional error design in the full adder array into the chip, the chip structure is simplified, the computing power consumption ratio optimization problem in the existing technology is solved, and the area and power consumption reduction is achieved, while maintaining the Efficient computing power performance.

CN112506471BActive Publication Date: 2024-11-15SHENZHEN MICROBT ELECTRONICS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011519912.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-21
Publication Date
2024-11-15
Estimated Expiration
2040-12-21

Smart Images

  • Figure CN112506471B_ABST
    Figure CN112506471B_ABST
Patent Text Reader

Abstract

The present invention discloses a chip and a computing system. A chip is provided, the chip comprising a plurality of registers for storing data; a logic operation module for performing operations; a fault-tolerant operation module associated with the registers and / or the logic operation module to receive a plurality of inputs, the plurality of inputs comprising a first input and a second input, the first input and the second input comprising the same number of bits, the fault-tolerant operation module comprising: an operation unit configured to obtain a sum output of a sum result of the plurality of inputs, and a carry unit configured to obtain a carry output of a sum result of the plurality of inputs; wherein the fault-tolerant operation module is configured to intentionally cause errors in its operation result for some input values ​​of the plurality of inputs.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to a chip and a computing system for computing. Background Art

[0002] Currently, chips based on the Proof of Work (POW) mechanism usually use hash tables as calculation content. Commonly, some chips use the SHA256 algorithm.

[0003] The focus is often on the following three aspects: the size of the chip, which determines the production cost; the speed of the chip, which determines the computing power, that is, the computing power; and the power consumption of the chip, which determines the degree of power consumption. The power consumption per unit of computing power (power consumption to computing power ratio) or the computing power power consumption ratio is an important indicator of performance.

[0004] Therefore, improving chip computing power, reducing chip production costs, and reducing chip power consumption are the main aspects of chip development in the industry.

[0005] Here, a new chip and computing system for computing are proposed. Summary of the invention

[0006] According to one aspect of the present disclosure, a chip is provided, comprising a fault-tolerant logic circuit, wherein the fault-tolerant logic circuit comprises: a plurality of registers for storing data; a logic operation module for performing operations; a fault-tolerant operation module associated with the registers and / or the logic operation module to receive a plurality of inputs, wherein the plurality of inputs comprises a first input and a second input, wherein the first input and the second input comprise the same number of bits; wherein the fault-tolerant operation module is configured to intentionally cause errors in its operation results for some input values ​​of the plurality of inputs.

[0007] In some embodiments, the fault-tolerant operation module includes: an operation unit configured to obtain a sum output of the sum results of the multiple inputs, and a carry unit configured to obtain a carry output of the sum results of the multiple inputs.

[0008] In some embodiments, the plurality of inputs further includes a carry input.

[0009] In some embodiments, the multiple inputs also include a third input having the same number of bits as the first and second inputs, and the carry unit is configured to perform the following logical operation on the bits A, B and C of the corresponding bits in the first, second and third inputs: Carry output = (A·B)+(B·C)+(A·C) The operation unit is configured to invert the carry output as a summed output, where "+" represents an OR operation and "·" represents an AND operation.

[0010] In some embodiments, the multiple inputs also include a third input, which has the same number of bits as the first and second inputs, and the operation unit is configured to perform the following logical operation on the bits A, B and C of the corresponding bits in the first, second and third inputs to obtain a sum output: Sum output = ~{(A·B)+(B·C)+(A·C)} The carry unit is configured to invert the sum output as the carry output, where "+" represents an OR operation, "·" represents an AND operation, and "~" represents an inverse operation.

[0011] In some embodiments, the plurality of inputs further comprises a third input having the same number of bits as the first and second inputs, the first, second and third inputs having corresponding bits A, B and C. In some embodiments, the fault-tolerant operation module is configured to intentionally make 2 of the logic operation truth table for A, B, C 3 In some embodiments, the operation unit is configured to intentionally make 2 of the logic operation truth tables for A, B, and C 3 At most 2 of the values ​​are in error.

[0012] In some embodiments, the logic operation module is configured to obtain data from one or more of the plurality of registers to perform logic operations thereon.

[0013] In some embodiments, the logic operation module includes a module adapted to perform one or more of Maj, Ch, and Σ operations as specified by the secure hash algorithm SHA-256.

[0014] In some embodiments, the fault-tolerant operation module includes a module adapted to perform one or more of Maj, Ch, and Σ operations as specified by the secure hash algorithm SHA-256.

[0015] In some embodiments, the chip includes a plurality of cores, each core includes the fault-tolerant logic circuit, and at least a portion of the plurality of cores are serially connected to each other.

[0016] In some embodiments, the chip further includes a top module that communicates with the multiple cores and verifies calculation results of the multiple cores.

[0017] In some embodiments, the chip is adapted to operate according to a secure hash algorithm SHA256 algorithm.

[0018] According to one aspect of the present disclosure, a computing system is provided, comprising the chip according to any embodiment.

[0019] In some embodiments, the computing system further includes a control board that communicates with the chip and verifies the computing results of the chip.

[0020] Other features and advantages of the present disclosure will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A block diagram schematically showing a calculation of the SHA256 algorithm used for operation;

[0022] Figure 2 A block diagram schematically illustrates a calculation of an improved SHA256 algorithm for operation;

[0023] Figure 3 A schematic diagram of the structure of a full adder array is shown schematically;

[0024] Figure 4 A schematic diagram showing the structure of a chip according to an exemplary embodiment of the present disclosure is shown;

[0025] Figure 5 A schematic diagram schematically showing the structure of a full adder;

[0026] Figure 6 Two different full adder implementations are schematically shown;

[0027] Figure 7 A circuit diagram schematically showing an XOR operation on three inputs;

[0028] Figure 8 A circuit diagram of a fault-tolerant adder according to an embodiment of the present disclosure is schematically shown;

[0029] Fig. 9 is a block diagram showing the fault-tolerant adder represented by a MAJ block;

[0030] Fig.10 A circuit diagram schematically showing another fault-tolerant adder according to an embodiment of the present disclosure; and

[0031] Fig.11 A logic operation truth table according to an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0032] Hereinafter, various exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. It should be understood that the following description of at least one exemplary embodiment is merely illustrative and is not intended to limit the present disclosure and its application or use. It should also be understood that any implementation described exemplarily herein does not necessarily mean that it is preferred or advantageous over other implementations. The present disclosure is not limited by any stated or implied theory given in the above technical field, background technology, content of the invention or specific implementation.

[0033] In addition, certain terms may be used in the following description for reference purposes only and are not intended to be limiting. For example, the words "first," "second," and other such numerical words referring to structures or elements do not imply a sequence or order unless the context clearly indicates otherwise.

[0034] It should also be understood that when the term “include / comprises” is used in this document, it indicates the presence of the specified features, integers, steps, operations, units and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, units and / or components and / or their combinations.

[0035] Figure 1 A calculation block diagram of a SHA256 algorithm used for operation is schematically shown. Figure 1 A to H shown in the figure represent registers (e.g., 32 bits), W t and K t Respectively represent data (for example, also 32 bits), and Maj, Ch, Σ1 and Σ0, for example, respectively represent the following logical operations:

[0036]

[0037] For more detailed descriptions of the above logic operations, see, for example, the description of SHA256 in the secure hash algorithm, which can be easily obtained by those skilled in the art from relevant websites. Some relevant descriptions are listed below by way of example.

[0038]

[0039]

[0040] Figure 2 A calculation block diagram of an improved SHA256 algorithm for operation is schematically shown. Figure 2A to H and R shown in the figure represent registers (for example, 32 bits), faa (Full adder array) represents a full adder array to be described later or may also be called a carry save adder, and cla (Carry lookaheadadder) represents a carry lookahead adder (which is a fast implementation of a full adder). Figure 1 Similar to that shown in , Maj, Ch, Σ1, Σ0, σ0 and σ1 represent the following logical operations respectively:

[0041]

[0042] For more detailed descriptions of the above logic operations, please refer to, for example, the description of SHA256 in the secure hash algorithm, which can be easily obtained by those skilled in the art from relevant websites.

[0043] Figure 3 The structure diagram of a full adder array is schematically shown. Figure 3 As shown, the full adder array includes a plurality of (here, four) full adders FA. Figure 3 In the example, a0 to a3 and b0 to b3 represent input values, c0 to c3 represent carry inputs, m0 to m3 represent sum outputs, and n1 to n4 represent carry outputs. Figure 3 The structure shown can, for example, implement a four-bit full adder.

[0044] Figure 4 A schematic structural diagram of a chip according to an exemplary embodiment of the present disclosure is shown. Figure 4 The chip shown in the figure includes a top module and multiple cores. The multiple cores can be arranged in a matrix form; however, the present disclosure is not limited thereto. It should be understood that the number of these cores is merely exemplary, and the chip according to the present disclosure may include more or fewer cores.

[0045] In some embodiments of the present disclosure, the multiple cores in the chip may be the same. For example, the multiple cores may be configured to perform the same operation on the same received signal. Alternatively, the multiple cores may be configured to perform calculations on the signals received by each based on the same algorithm. Alternatively, the multiple cores may be configured to perform calculations on the signals received by each based on the same algorithm. In some embodiments, the cores may calculate the data received by each based on the same known algorithm or an algorithm to be developed in the future, for example, the SHA256 algorithm or the improved SHA256 algorithm used for calculations as described above, and feed back the valid calculation results calculated by each to the top module. The top module may verify the valid calculation results from all modules (or cores), and then further feed back the correct valid calculation results to the verification module at the upper level (for example, the control board, not shown in the figure). There may also be multiple verification checks at the upper level.

[0046] During the calculation process, the top module or control board will verify the results; if the core calculation is wrong and the calculation results are reported to the top module or control board by mistake, the top module or control board will discard the wrong results. This will not have a significant impact on the correctness of the results or the probability of obtaining the final calculation results. Figure 3 As shown in , the number of cores is large and the calculation is intensive; the top module will verify the calculation results reported by the core. Therefore, it would be desirable if the core structure could be simplified to reduce the production cost and operating power consumption (which can be regarded as the operating cost) of the chip, while at the same time having little or no effect on the probability of obtaining the expected result, thereby improving the computing power consumption ratio.

[0047] To this end, the inventor of the present application has proposed a chip for computing disclosed herein. The chip disclosed herein adopts a fault-tolerant mechanism, which simplifies the chip structure and can significantly reduce the power consumption and size of the chip, while ensuring that the chip performance does not decrease too much, so as to improve (enhance) the computing power consumption ratio of the chip. Through fault-tolerant design, it is possible to substantially save area and reduce power consumption, and maintain or improve the computing power cost ratio.

[0048] Next, the chip for computing of the present disclosure will be described in detail with reference to specific examples.

[0049] According to an embodiment of the present disclosure, a chip may include a fault-tolerant logic circuit, which may include: a plurality of registers for storing data, a logic operation module for performing operations, and one or more fault-tolerant operation modules. In some implementations, the fault-tolerant logic circuit may also be included in one or more cores.

[0050] According to some embodiments of the present disclosure, the plurality of registers may be as described above in combination with Figure 1 The register AH shown above can also be combined with Figure 2 The registers AH and R are shown. The registers AH can store data, for example, data received from the outside or intermediate data used internally.

[0051] The logic operation module can perform logic operations. According to some embodiments of the present disclosure, the logic operation module can obtain data from one or more registers of the plurality of registers to perform logic operations on the data. For example, the logic operation module can perform the above combined Figure 1 and Figure 2 One or more of the Maj, Ch, Σ1, Σ0, σ0 and σ1 operations described above; however, the present disclosure is not limited thereto. In some embodiments, multiple logic operation modules may be included.

[0052] According to some embodiments of the present disclosure, the fault-tolerant operation module is configured to intentionally cause errors in its operation results for some of the input values ​​of the multiple inputs. Unlike the related art, which generally expects to perform operations comprehensively and accurately and avoid errors, the inventors of the present application have proposed a fault-tolerant technology that intentionally introduces errors. Through the technology disclosed in the present disclosure, the circuit structure of the chip can be simplified, the chip area can be reduced, the power consumption can be reduced, and / or the cost per unit of computing power can be reduced, or the ratio of computing power to cost (for example, the computing power to power consumption ratio) can be improved. In addition, according to the technology disclosed in the present disclosure, there is almost no effect on the probability of obtaining the desired result, while significantly reducing the cost.

[0053] Next, the fault-tolerant operation module according to the present disclosure will be described in detail.

[0054] In some embodiments of the present disclosure, the fault-tolerant design of the chip is performed for the full adder faa to simplify the circuit structure of the faa module, thereby reducing the area and power consumption of the chip. However, it should be understood that those skilled in the art can apply its principles to other modules based on the teachings of this application. For example, the fault-tolerant design can also be performed for the carry-lookahead adder cla module; or, the fault-tolerant design can also be performed for one or more of the operation modules such as Maj, Ch, and Σ specified by the secure hash algorithm SHA-256; however, it should be understood that the present disclosure is not limited to this.

[0055] First, a full adder will be described in detail. Figure 5 The schematic diagram shows the structure of a full adder. Figure 5 As shown in, for example, A and B represent binary inputs, and C in represents carry input, Sum represents sum output, and C out Indicates carry output (also called carry). Usually, the following logical operation is performed for a one-bit full adder:

[0056] … (1)

[0057] C out =(A·B) +[C in ·(A+B)] = (A·B) + (B·C in ) + (A·C in ) …(2),

[0058] in," " represents an exclusive OR operation, "+" represents an OR operation, and "·" represents an AND operation. It should be understood that although the above-mentioned symbolic representation is basically used to represent logical operations in the context of this article, other symbolic representations can also be used (not excluded).

[0059] Figure 6 Two different implementations of full adders are schematically shown. Figure 6 Figure (a) shows a standard CMOS implementation of a full adder, and Figure 6 Figure (b) in shows the multiplex implementation of the full adder. Similarly, Figure 6 a i , b i Represents a binary input, C in Indicates carry input, S i represents the summation output, and C out Since the above-mentioned full adder implementation is known in the art, its details will not be described in more detail here.

[0060] Figure 7 A three-input full adder implementation is shown, in which the three inputs are XORed. It is easy to understand that in some cases, C can also be input with a carry; but it is not limited to this. Figure 7 In the example, the following logical operations are performed:

[0061] Sum output Sum=A B C… (3),

[0062] Carry output (not shown) = (A·B) +[C·(A+B)].

[0063] Wherein, A, B and C represent a binary input respectively. In some embodiments, the following Figure 7 In some embodiments, there is a situation where the basic structural unit of three tubes in series is not allowed to be used due to low voltage, then it is necessary to Figure 7The circuit shown for performing an XOR operation on three inputs is split into two two-input XOR gates (XOR2) connected in series. In this case, the following logic operation is performed:

[0064] SUM= (A B) C… (4),

[0065] In this case, the logic circuit structure will become like Figure 7 The circuit structure shown is complex.

[0066] Figure 8 A circuit diagram of an adder (herein referred to as a fault-tolerant adder) designed based on a full adder according to an embodiment of the present disclosure is schematically shown. Fig. 9 The figure shows a logic block diagram of the fault-tolerant adder using the aforementioned MAJ module.

[0067] As mentioned above, for a three-input full adder receiving inputs A, B, and C, its carry output Carry = (A·B) +[C·(A+B)], and its sum output Sum = A B C. Here, the inventor of the present application conceives to introduce an error into the full adder to simplify its structure.

[0068] As a non-limiting example, the inventors set the sum output to: Sum = ~ Carry, that is, the sum output is the inverse of the carry output, where "~" represents inversion. Therefore, under this setting, for the three inputs A, B and C, the calculation of the carry output Carry is accurate, that is, the 8 outputs of the 8 input situations of the truth table of the three inputs A, B and C are accurate, that is, the accuracy rate is 8 / 8. For the sum output Sum, 2 of the 8 outputs of the 8 input situations of the truth table of the three inputs A, B and C are wrong, and the remaining 6 are accurate; that is, the accuracy rate is 6 / 8. For the overall sum result, the accuracy rate is 14 / 16.

[0069] Furthermore, in this setting, the carry operation corresponds exactly to the aforementioned MAJ(A, B, C) operation. That is,

[0070] Carry=MAJ(A,B,C)=(A·B)+(B·C)+(A·C)…(5).

[0071] Typically, chip manufacturers will provide a standard unit MAJ, or also known as the AO222 unit (the name or model may vary depending on the manufacturer).

[0072] Since the sum output is set to Sum = ~ Carry, Sum = ~MAJ(A, B, C) = MAJI(A, B, C). Here, MAJI means the inverse of MAJ. Figure 8 In the example, the portion on the left side of the node outputting the summation result Sum can be represented as MAJI, and the portion on the right side thereof can be represented as INV, where INV represents inversion of the input result (for example, it can be implemented with an inverter (such as a NOT gate)).

[0073] Thus, the fault-tolerant adder can be represented by a MAJ module and an INV gate, as Fig. 9 As shown. Fig. 9 In the example, the output of MAJ is the carry output Carry, that is, Carry = MAJ (A, B, C), and the output of MAJ is obtained by inverter INV to obtain Sum, that is, Sum = ~MAJ (A, B, C) = MAJI (A, B, C).

[0074] contrast Figure 8 and Figure 7 The structure shown is as follows: Figure 7 The full adder shown in FIG. 1 only includes a circuit for summing output and does not include a circuit for carrying output. However, even in this case, it still includes 16 transistors. In contrast, Figure 8 As shown, the fault-tolerant adder according to the embodiment of the present application only has 12 transistors when including the circuit part for carry output.

[0075] Therefore, according to the technology disclosed in the present invention, the chip area can be significantly reduced. In addition, compared with the full adder of the prior art, the technology according to the present invention substantially reduces the XOR gate and the transistor, thereby reducing power consumption, reducing the requirements for circuit performance or other parameters, and increasing the degree of freedom of design.

[0076] It should be understood that, although the above description is made with an input of one bit as an example, it is obvious to those skilled in the art that the principle is applicable to the case where the input has multiple bits. In fact, for inputs of multiple bits (e.g., 32 bits or more), the area of ​​the chip can be significantly reduced according to the present invention.

[0077] It is easy to understand that the input of the above-mentioned fault-tolerant adder can be received from the aforementioned register and / or other operation modules.

[0078] At this point, it should be understood that, according to an embodiment of the present disclosure, the fault-tolerant operation module can be associated with a register and / or a logic operation module to receive multiple inputs. These multiple inputs may include a first input and a second input, and the first input and the second input may include the same number of bits. The fault-tolerant operation module may include an operation unit and a carry unit. The operation unit is configured to obtain a sum output of the summation result of the multiple inputs, and the carry unit is configured to obtain a carry output of the summation result of the multiple inputs. According to some embodiments of the present disclosure, the multiple inputs also include a carry input. According to some embodiments of the present disclosure, the multiple inputs also include a third input, and the third input has the same number of bits as the first input and the second input.

[0079] Fig.10 A circuit diagram of another adder based on a full adder for fault tolerance design according to an embodiment of the present disclosure is schematically shown. According to this embodiment of the present disclosure, due to the following logical operations:

[0080] (A+B) ·[(A·B)+C] =[(A+B) ·(A·B)] +[(A+B)·C]= (A·B) +[(A+B)·C],

[0081] therefore, Fig.10 The circuit shown is logically equivalent to Figure 8 The circuit shown can be realized according to Figure 8 The logical functions described above are the same as the logical functions described above. Therefore, they will not be described in detail here. In other words, different circuit implementations can be used to achieve the same or equivalent logical functions.

[0082] Fig.11 Schematically shows a logic operation truth table of a fault-tolerant adder according to an exemplary embodiment of the present disclosure. Fig.11 As shown, Ai represents the addend (e.g., the i-th bit), Bi represents the addend (e.g., the i-th bit), Ci-1 represents the carry number from the adjacent low bit (e.g., the i-1th bit), Si represents the sum of the output bit (or called the current bit, e.g., the i-th bit), and Ci represents the carry number to the adjacent high bit (e.g., the i+1th bit). According to some embodiments of the present disclosure, the operation unit in the fault-tolerant operation module is configured to intentionally cause errors in at most 2 values ​​of the output Si in the logic operation truth table for Ai, Bi, and Ci-1. As shown in the figure, the calculation result of Ci (i.e., the carry) is completely correct, while the calculation result for Si (i.e., Sum) has errors of at most 2 values.

[0083] In this way, although the calculation results of the fault-tolerant calculation module are intentionally made to have errors for some input values ​​of multiple inputs, so that 14 of the 16 output values ​​are correct, that is, the accuracy rate is 14 / 16, the circuit structure can be greatly simplified as described above.

[0084] In some embodiments, the chip may include multiple cores, each core including the fault-tolerant logic circuit described above. In some embodiments, the chip further includes a top module that communicates with the multiple cores and verifies the calculation results of the multiple cores.

[0085] Consider some application scenarios according to the embodiments of the present application below. For example, when applied to the calculation of the SHA256 algorithm, since SHA256 itself can tolerate errors to a certain extent, the computing cost will be further reduced. For example, consider the CH operation module in the SHA256 algorithm. CH (x, y, z) = (x AND y) XOR ((NOT x) AND z), where x, y, and z represent input data respectively, where XOR represents exclusive OR, AND represents AND, and NOT represents NOT. When y = z, no matter what the value of x is, the output result is the same, that is, even if x has an error, it will not affect the correctness of the calculation. For another example, considering MAJ (x, y, z) = (x AND y) XOR (xAND z) XOR (y AND z), when x and y are the same, no matter what the value of z is, the output result is the same, that is, even if z has an error, it will not affect the correctness of the calculation. Therefore, when the fault-tolerant module of the embodiment of the present disclosure is used, even if some output results have errors, when they are provided to the MAJ module or the CH module as input, in some cases, the errors do not affect the accuracy of the CH or MAJ operation. Therefore, the impact on the final operation result is limited, and on the contrary, the manufacturing cost and power consumption are greatly reduced.

[0086] In addition, as mentioned above, in some applications, the upper module will also verify the valid final calculation results. For example, when a calculation error occurs in the fault-tolerant logic circuit in the chip (which can be included in the core in the chip), such as calculating the correct calculation result incorrectly or failing to calculate the correct calculation result, the core will feed back the wrong calculation result to the top module or control board. However, since the top module or control board will verify the calculation results fed back by multiple cores in the chip, the wrong calculation results will be discarded, so that the correctness of the final operation of the chip will not be basically affected. For example, the control board can communicate with the chip and verify the calculation results of the chip.

[0087] Therefore, according to the technology of the present application, the impact on the probability of obtaining the expected computing results is minimal, while the cost and power consumption are significantly reduced, thereby greatly reducing the cost of computing power and improving the computing power to power consumption ratio.

[0088] Those skilled in the art will appreciate that the description of the boundaries between operations (or steps) in the above embodiments is merely illustrative. Multiple operations may be combined into a single operation, a single operation may be distributed in additional operations, and operations may be performed at least partially overlapping in time. Moreover, alternative embodiments may include multiple instances of a particular operation, and the order of operations may be changed in various other embodiments. However, other modifications, variations, and substitutions are equally possible. Therefore, this specification and the accompanying drawings should be viewed as illustrative, not restrictive.

[0089] Although some specific embodiments of the present disclosure have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present disclosure. The various embodiments disclosed herein may be combined in any manner without departing from the spirit and scope of the present disclosure. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A chip suitable for performing operations according to the secure hash algorithm SHA256 algorithm, It is characterized in that A fault-tolerant logic circuit is included, wherein the fault-tolerant logic circuit includes: Multiple registers for storing data; A logic operation module, used for performing operations; a fault-tolerant operation module, the fault-tolerant operation module comprising a fault-tolerant adder, the fault-tolerant adder being associated with the register and / or the logic operation module to receive a plurality of inputs, the plurality of inputs comprising a first input, a second input and a third input, the first input, the second input and the third input comprising the same number of bits, Wherein, the fault-tolerant adder comprises: an operation unit configured to obtain a sum output of the sum results of the plurality of inputs, and a carry unit configured to obtain a carry output of the summation result of the plurality of inputs, One or both of the operation unit and the carry unit are configured to intentionally cause errors to appear in the operation results of the error-tolerant adder for some input values ​​of the multiple inputs.

2. The chip according to claim 1, It is characterized in that The plurality of inputs also includes a carry input.

3. The chip according to claim 1, It is characterized in that The carry unit is configured to perform the following logic operation on the bits A, B and C of the corresponding bits in the first, second and third inputs: Carry out = (A·B)+(B·C)+(A·C) The arithmetic unit is configured to invert the carry output as the sum output, Here, "+" represents an OR operation, and "·" represents an AND operation.

4. The chip according to claim 1, It is characterized in that The plurality of inputs further includes a third input having the same number of bits as the first and second inputs, The operation unit is configured to perform the following logic operation on the bits A, B and C of the corresponding bits in the first, second and third inputs to obtain a sum output: Sum output = ~{(A·B)+(B·C)+(A·C)} The carry unit is configured to invert the sum output as a carry output, Among them, "+" represents the OR operation, "·" represents the AND operation, and "~" represents the inverse operation.

5. The chip according to claim 1, It is characterized in that The first, second and third inputs have bits A, B and C of corresponding positions, The fault-tolerant adder is configured to intentionally make 2 of the truth tables for the logic operations of A, B, and C 3 At most 2 of the values ​​are in error.

6. The chip according to claim 1, It is characterized in that The logic operation module is configured to obtain data from one or more of the plurality of registers to perform logic operations thereon.

7. The chip according to claim 1, It is characterized in that The logic operation module includes a module adapted to perform one or more of the Maj, Ch, Σ0 and Σ1 operations specified in the secure hash algorithm SHA-256: The Maj, Ch, Σ0 and Σ1 operations are as follows: Where x, y and z are operands respectively.

8. The chip according to claim 1, It is characterized in that The chip includes a plurality of cores, each core includes the fault-tolerant logic circuit, and at least a portion of the plurality of cores are serially connected to each other.

9. The chip according to claim 8, It is characterized in that The chip also includes a top module, which communicates with the multiple cores and verifies the calculation results of the multiple cores.

10. The chip according to claim 1, It is characterized in that The fault-tolerant operation module includes a module adapted to perform one or more of the Maj, Ch, Σ0 and Σ1 operations specified in the secure hash algorithm SHA-256, The Maj, Ch, Σ0 and Σ1 operations are as follows: Where x, y and z are operands respectively.

11. A computing system, It is characterized in that Comprising a chip according to any one of claims 1-10.

12. The computing system according to claim 11, It is characterized in that It also includes a control board, which communicates with the chip and verifies the calculation results of the chip.