KR20210085176A - Computing device capable of code-based data operation and operating method thereof - Google Patents

Abstract

Disclosed are a computing device capable of processing data arithmetic based on a sign and an operating method thereof. According to the present invention, when a data arithmetic process is performed in the computing device, the data arithmetic process is performed based on a coded message after encoding a message, rather than using a plaintext message as it is, thereby improving security in the data arithmetic process. According to the present invention, the computing device (110) includes a generation matrix storage unit (111), an encoding unit (112), a data operation unit (113), and a decoding unit (114).

Description

COMPUTING DEVICE CAPABLE OF CODE-BASED DATA OPERATION AND OPERATING METHOD THEREOF

The present invention relates to a computing device capable of performing sign-based data arithmetic processing and an operating method thereof.

Recently, as computing devices are utilized in various fields, a security technology capable of preparing for attacks on computing devices has emerged as an important issue.

In particular, as computing devices are recently introduced to large-scale industrial facilities such as nuclear power plants or autonomous vehicles, it is essential to introduce intelligent security technology capable of defending against malicious attacks from the outside.

The computing device performs data operation through various operation processing modules. If malicious data manipulation or error generation from the outside occurs in the process of data calculation inside the computing device, a big problem may occur due to data calculation being incorrectly processed.

Data operation processing in the existing computing device has a problem in that it is vulnerable to data manipulation from the outside in that it uses the plaintext message generated by the computing device as it is. For example, when it is assumed that an operation of ORing the first message and the second message generated by the computing device is performed, the existing computing device logically ORs the first message and the second message in a plaintext message state, so that the third There was a problem of being easily exposed to data manipulation by a person.

Therefore, when data arithmetic processing is performed in the computing device, the security in the data arithmetic process can be strengthened by encoding the message and then performing the data arithmetic processing based on the coded message, rather than using the plaintext message as it is. Research on existing technologies is needed.

When data operation processing is performed in a computing device, the present invention does not utilize a plaintext message as it is, but encodes the message and then enables data operation processing to be performed based on the coded message, thereby enhancing security in the data operation process. make it possible

A computing device capable of performing sign-based data arithmetic processing according to an embodiment of the present invention has a preset kxn (k and n are natural numbers of 2 or more, n is k) having an error correction capability for t (t is a natural number) bit code. When a generation matrix storage unit in which a generator matrix of a larger size is stored, and a first operation command to perform an OR operation on the first message and the second message generated by the computing device is applied, the A first codeword is generated by encoding the first message based on a generation matrix, and a second codeword is generated by encoding the second message based on the generation matrix. an encoding unit that generates a first operation codeword by ORing the first codeword and the second codeword when the first codeword and the second codeword are generated, and then decodes the first operation codeword The first message and the first message are received by receiving the first operation codeword from the data operation unit and the data operation unit transmitted to the unit, and performing error correction-based decoding on the first operation codeword based on the generation matrix. and the decoding unit for restoring the sum message in which the two messages are ORed.

In addition, according to an embodiment of the present invention, the method of operating a computing device capable of performing sign-based data operation processing includes a preset kxn (k and n is 2 or more) having an error correction capability for t (t is a natural number) bit code. Maintaining a generation matrix storage in which a generation matrix of a size (n is a number greater than k) is stored as a natural number, and a first operation instruction for performing an OR on the first message and the second message generated by the computing device. When applied, the encoding unit generates a first codeword by performing encoding on the first message based on the generation matrix, and performs encoding on the second message based on the generation matrix to obtain a second code generating a word; when the first codeword and the second codeword are generated, a data operation unit generates a first operation codeword by ORing the first codeword and the second codeword, and then the first operation codeword is generated. transmitting an operation codeword to a decoding unit, and the decoding unit receives the first operation codeword from the data operation unit, and performs error correction-based decoding on the first operation codeword based on the generation matrix, and restoring a sum message in which the first message and the second message are ORed together.

When data operation processing is performed in a computing device, the present invention does not utilize a plaintext message as it is, but encodes the message and then enables data operation processing to be performed based on the coded message, thereby enhancing security in the data operation process. can

1 is a diagram illustrating a structure of a computing device capable of performing code-based data arithmetic processing according to an embodiment of the present invention.
2 is a flowchart illustrating a method of operating a computing device capable of performing sign-based data operation processing according to an embodiment of the present invention.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. These descriptions are not intended to limit the present invention to specific embodiments, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. While describing each drawing, like reference numerals are used for similar components, and unless otherwise defined, all terms used in this specification, including technical or scientific terms, refer to those of ordinary skill in the art to which the present invention belongs. It has the same meaning as is commonly understood by those who have it.

In this document, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. In addition, in various embodiments of the present invention, each of the components, functional blocks or means may be composed of one or more sub-components, and the electrical, electronic, and mechanical functions performed by each component are electronic. A circuit, an integrated circuit, an ASIC (Application Specific Integrated Circuit), etc. may be implemented with various well-known devices or mechanical elements, and may be implemented separately or two or more may be integrated into one.

On the other hand, the blocks in the accompanying block diagram or steps in the flowchart are computer program instructions that are loaded in a processor or memory of equipment capable of data processing, such as a general-purpose computer, a special-purpose computer, a portable notebook computer, and a network computer, and perform specified functions. can be interpreted as meaning Since these computer program instructions may be stored in a memory provided in a computer device or in a memory readable by a computer, the functions described in the blocks of the block diagrams or the steps of the flowcharts are produced as articles of manufacture containing instruction means for performing the same. could be In addition, each block or each step may represent a module, segment, or portion of code comprising one or more executable instructions for executing the specified logical function(s). It should also be noted that, in some alternative embodiments, it is also possible for the functions recited in blocks or steps to be executed out of the prescribed order. For example, two blocks or steps shown one after another may be performed substantially simultaneously or in the reverse order, and in some cases, some blocks or steps may be omitted.

1 is a diagram illustrating a structure of a computing device capable of performing code-based data arithmetic processing according to an embodiment of the present invention.

Referring to FIG. 1 , the computing device 110 according to the present invention includes a generation matrix storage unit 111 , an encoding unit 112 , a data operation unit 113 , and a decoding unit 114 .

In the generator matrix storage unit 111, a generator matrix having a preset size of kxn (k and n are natural numbers greater than or equal to 2 and n is a number greater than k) having error correction capability for t (t is a natural number) bit code. ) is stored.

The generation matrix is a matrix used for generating a codeword so that the data receiving side can perform error correction when the data transmitting side transmits data to the data receiving side, and the data transmitting side is based on a predetermined generation matrix When the codeword is generated and transmitted to the data receiving side, the data receiving side can restore the error and the original data from the received data through the parity check matrix corresponding to the generation matrix.

The encoding unit 112 encodes the first message based on the generation matrix when a first operation command for performing a logical OR on the first message and the second message generated by the computing device 110 is applied. By performing, the first codeword is generated, and the second codeword is generated by encoding the second message based on the generation matrix.

For example, if the generation matrix is 'G', the message vector for the first message is 'm ₁ ', and the message vector for the second message is 'm ₂ ', the encoding unit 112 is _{The first codeword 'c 1} ' may be generated according to the operation of Equation 1, _{and the second codeword 'c 2} ' may be generated according to the operation of Equation 2 below.

When the first codeword and the second codeword are generated, the data operation unit 113 generates a first operation codeword by ORing the first codeword and the second codeword, and then the first operation codeword is transmitted to the decoding unit.

In relation to this, the data operation unit 113 may generate the first operation codeword 'c ₁ +c ₂ ' according to the operation of Equation 3 below.

The decoding unit 114 receives the first operation codeword from the data operation unit 113 and performs error correction-based decoding on the first operation codeword based on the generation matrix, thereby generating the first message and The sum message in which the second message is ORed is restored.

At this time, according to an embodiment of the present invention, the decoding unit 114 generates a parity check matrix corresponding to the generation matrix, and syndrome decoding for the first operation codeword based on the parity check matrix. ), it is possible to restore the summed message in which the first message and the second message are ORed together.

First, the parity check matrix is a matrix for verifying whether an error occurs in the data received by the data receiving side when the data transmitting side transmits a codeword encoded based on a predetermined generation matrix to the data receiving side, kxn The size of the parity check matrix 'H' corresponding to the generation matrix 'G' of the size is (nk) x n, and has characteristics as shown in Equation 4 below in relation to the generation matrix 'G'.

Here, H ^T denotes a transpose matrix of the parity check matrix.

In relation to this, when the first operation codeword is transmitted from the data operation unit 113, the decoding unit 114 generates a parity check matrix corresponding to the generation matrix, and then applies the parity check matrix to the first operation codeword. An operation according to Equation 5 below may be performed based on .

Here, 'r' is the first operation codeword transmitted from the data operation unit 113, and the first operation codeword is '(m ₁ +m ₂ )G', but data A predetermined error may be applied in the operation process or the data transfer process. If an error vector according to an error applied to the first operation codeword is 'e', Equation 5 may be expressed as Equation 6 below.

는 배타적 논리합 연산을 의미한다. 상기 수학식 5에서와 같은 연산을 통해 산출된 'S'를 신드롬(Syndrome)이라고 하는데, 상기 수학식 5에 따라 연산되는 신드롬 'S'는 상기 수학식 6에서 나타낸 바와 같이,

와 동일하기 때문에, 디코딩부(114)는 상기 수학식 5의 연산을 통해 신드롬 'S'를 산출한 후 상기 신드롬 'S'와 상기 패리티 검사 행렬을 기초로

를 만족하는 에러 벡터 'e'를 찾아낼 수 있다.here,

is an exclusive-OR operation. 'S' calculated through the operation as in Equation 5 is called a syndrome, and the syndrome 'S' calculated according to Equation 5 is as shown in Equation 6,

Since it is the same as , the decoding unit 114 calculates the syndrome 'S' through the operation of Equation 5, and then based on the syndrome 'S' and the parity check matrix.

It is possible to find an error vector 'e' that satisfies .

_{Then, the decoding unit 114 restores '(m 1} +m ₂ )G', which is the first operation codeword from which the error is removed, based on the error vector 'e', and from the first operation codeword _{'(m 1} +m ₂ )', which is the sum message of the first message and the second message, may be restored.

According to an embodiment of the present invention, the encoding unit 112 receives a second operation command to perform convolution on the third message and the fourth message generated by the computing device 110 is applied. A third codeword may be generated by performing encoding on the third message based on a generation matrix, and a fourth codeword may be generated by performing encoding on the fourth message based on the generation matrix. .

For example, if the generation matrix is 'G', the message vector for the third message is 'm ₃ ', and the message vector for the fourth message is 'm ₄ ', the encoding unit 112 is _{The third codeword 'c 3} ' may be generated according to the operation of Equation 7, _{and the fourth codeword 'c 4} ' may be generated according to the operation of Equation 8 below.

At this time, when the third codeword and the fourth codeword are generated, the data operation unit 113 generates a second operation codeword by performing Hadamard product of the third codeword and the fourth codeword. The second operation codeword may be multiplied by a preset bootstrap matrix to convert the second operation codeword into a convolutional codeword, and then the convolutional codeword may be transmitted to the decoding unit 114 .

Here, the convolutional codeword refers to a codeword obtained by performing encoding on the result of convolution of the third message and the fourth message based on the generation matrix.

In addition, the Hadamard product means the operation of multiplying each component in a vector or matrix of the same size. When the vectors [abc] and [xyz] exist, the Hadamard product between the two vectors is calculated as [ax by cz]. can

In relation to this, the data operation unit 113 may _{convert the second operation codeword into the convolution codeword 'c cov} ' according to the operation of Equation 9 below.

'는 아다마르곱을 의미한다. 상기 부트스트랩 행렬은 상기 수학식 9를 통해 연산되는 결과가 상기 제3 메시지와 상기 제4 메시지를 합성곱한 결과에 대해 상기 생성행렬을 기초로 인코딩을 수행한 코드워드와 같아지도록 미리 설계된 행렬이기 때문에, 상기 수학식 9에서 연산된 합성곱 코드워드는 하기의 수학식 10과 같이, 상기 제3 메시지와 상기 제4 메시지를 합성곱한 결과에 대해 상기 생성행렬을 기초로 인코딩을 수행한 코드워드와 같아지게 된다.Here, 'B' is the bootstrap matrix, '

' means Hadamarch. Since the bootstrap matrix is a matrix previously designed so that the result calculated through Equation 9 is equal to the codeword obtained by performing encoding on the result of convolution of the third message and the fourth message based on the generation matrix. , the convolution codeword calculated in Equation 9 is the same as the codeword obtained by performing encoding on the result of convolution of the third message and the fourth message based on the generation matrix as shown in Equation 10 below. will lose

'은 합성곱을 의미한다.here, '

' means convolution.

In this case, according to an embodiment of the present invention, the generation matrix stored in the generation matrix storage unit 111 may be a generation matrix composed of a primary Reed-Muller (RM) code.

'와 같이, 상기 제2 연산 코드워드로의 연산이 수행됨에 따라 2차 RM 코드로 변환된다.At this time, as the third codeword and the fourth codeword are Hadamard multiplied, the primary RM code applied to the third codeword and the fourth codeword, respectively, is a secondary RM in the second operation codeword. converted to code. That is, in Equations 7 and 8, the primary RM code constituting the generation matrix 'G' applied to the third codeword and the fourth codeword is '

', as the operation on the second operation codeword is performed, it is converted into a secondary RM code.

At this time, the bootstrap matrix converts the secondary RM code applied to the second operation codeword into the primary RM code and at the same time, in the generation matrix according to the converted primary RM code in the second operation codeword. The multiplied message value may be a matrix set to be the same as a result of convolution of the third message and the fourth message.

'라는 상기 제2 연산 코드워드가 생성되었을 때, 상기 제2 연산 코드워드를 소정의 메시지 'm_k'와 1차 RM 코드에 따른 생성행렬의 곱의 형태인 '

'로 변환하되, 메시지 'm_k'가 상기 제3 메시지와 상기 제4 메시지를 합성곱한 결과인 '

'와 동일해지도록 미리 설정된 행렬일 수 있다.That is, the bootstrap matrix is '

When the second operation codeword ' is generated, the second operation codeword is in _{the form of a product of a predetermined message 'm k} ' and a generation matrix according to the primary RM code.

', but the message 'm _k ' is the result of convolution of the third message and the fourth message.

It may be a matrix preset to be equal to '.

In this way, when the data operation unit 113 generates the convolution codeword, the decoding unit 114 receives the convolution codeword from the data operation unit 113 and based on the generation matrix for the convolution codeword. By performing error-correction-based decoding, it is possible to recover a convolutional message in which the third message and the fourth message are convolutioned.

At this time, according to an embodiment of the present invention, the decoding unit 114 generates a parity check matrix corresponding to the generation matrix, and performs syndrome decoding on the convolutional codeword based on the parity check matrix, The convolutional message obtained by convolution of the third message and the fourth message may be recovered.

In relation to this, when the convolution codeword is transmitted from the data operation unit 113, the decoding unit 114 generates a parity check matrix corresponding to the generation matrix, and then generates the parity check matrix for the convolution operation codeword. An operation according to the following Equation 11 may be performed based on the equation (11).

'이지만 데이터 연산 과정이나 데이터 전달 과정에서 소정의 에러가 적용될 수 있다. 만약, 상기 합성곱 코드워드에 적용된 에러에 따른 에러 벡터를 'e₁'이라고 하는 경우, 상기 수학식 11는 하기의 수학식 12와 같이 나타낼 수 있다.Here, 'r _cov ' is the convolution codeword transmitted from the data operation unit 113, and the convolution codeword is, as originally shown in Equation 10, '

' However, a certain error may be applied in the data operation process or data transfer process. If the error vector according to the error applied to the convolutional codeword is 'e ₁ ', Equation 11 can be expressed as Equation 12 below.

를 만족하는 에러 벡터 'e₁'를 찾아낼 수 있다.In this case, the decoding unit 114 calculates the syndrome 'S _cov ' through the operation of Equation (11), and then based on the syndrome 'S _cov ' and the parity check matrix.

It is possible to find an error vector 'e _{1 ' that satisfies .}

'을 복원하고, 상기 합성곱 코드워드로부터 상기 제3 메시지와 상기 제4 메시지가 합성곱된 합성곱 메시지인 '

'를 복원할 수 있다.Then, the decoding unit 114 ' is the convolutional codeword from which the error has been removed based on _{the error vector 'e 1 '.}

' is a convolutional message in which the third message and the fourth message are convolved from the convolution codeword.

' can be restored.

2 is a flowchart illustrating a method of operating a computing device capable of performing sign-based data operation processing according to an embodiment of the present invention.

In step S210, a generation matrix having a preset size of kxn (k and n are natural numbers greater than or equal to 2, and n is a number greater than k) having an error correction capability for t (t is a natural number) bit code is stored. maintain storage.

In step S220, when a first operation command for performing an OR operation on the first message and the second message generated by the computing device is applied, the encoding unit encodes the first message based on the generation matrix. By doing so, a first codeword is generated, and encoding is performed on the second message based on the generation matrix to generate a second codeword.

In step S230, when the first codeword and the second codeword are generated, a data operation unit generates a first operation codeword by ORing the first codeword and the second codeword, and then the first operation The codeword is transmitted to the decoding unit.

In step S240, the decoding unit receives the first operation codeword from the data operation unit, and performs error correction-based decoding on the first operation codeword based on the generation matrix, so that the first message and The sum message in which the second message is ORed is restored.

At this time, according to an embodiment of the present invention, in step S240, the decoding unit generates a parity check matrix corresponding to the generation matrix, and performs syndrome decoding for the first operation codeword based on the parity check matrix. By doing so, it is possible to recover the summed message in which the first message and the second message are ORed together.

In addition, according to an embodiment of the present invention, in the method of operating the computing device, when a second operation command for performing convolution on the third message and the fourth message generated by the computing device is applied, the encoding unit generating a third codeword by performing encoding on the third message based on a generation matrix, and generating a fourth codeword by performing encoding on the fourth message based on the generation matrix; When the third codeword and the fourth codeword are generated, the data operation unit generates a second operation codeword by Hadamard multiplication of the third codeword and the fourth codeword, and is added to the second operation codeword. The second operation codeword is multiplied by a preset bootstrap matrix for a convolutional codeword (the convolutional codeword is a result of convolution of the third message and the fourth message, encoding is performed based on the generation matrix) after converting to a single codeword), transferring the convolution codeword to the decoding unit, and the decoding unit receiving the convolution codeword from the data operation unit, and the generation matrix for the convolution codeword The method may further include reconstructing a convolutional message in which the third message and the fourth message are convolved by performing error correction-based decoding based on .

At this time, according to an embodiment of the present invention, the generation matrix is a generation matrix composed of a primary RM code, and the bootstrap matrix is a secondary RM code applied to the second operation codeword (the third codeword and As the fourth codeword is multiplied by Hadamard, the primary RM code applied to the third codeword and the fourth codeword is converted into a secondary RM code in the second operation codeword) A matrix set such that a message value multiplied by a generation matrix according to the converted primary RM code in the second operation codeword at the same time as conversion into an RM code becomes the same as a result of convolution of the third message and the fourth message can be

In addition, according to an embodiment of the present invention, in the step of restoring the convolution message, the decoding unit generates a parity check matrix corresponding to the generation matrix, and based on the parity check matrix, By performing syndrome decoding, it is possible to recover the convolutional message in which the third message and the fourth message are convolutioned.

In the above, an operation method of a computing device capable of performing a code-based data operation processing according to an embodiment of the present invention has been described with reference to FIG. Here, the operating method of a computing device capable of processing data based on a sign according to an embodiment of the present invention corresponds to the configuration and operation of the computing device 110 capable of processing processing for data based on a sign described with reference to FIG. 1 . Therefore, a more detailed description thereof will be omitted.

The method of operating a computing device capable of performing code-based data arithmetic processing according to an embodiment of the present invention may be implemented as a computer program stored in a storage medium for execution through combination with a computer.

In addition, the operating method of the computing device capable of performing code-based data arithmetic processing according to an embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, the present invention has been described with specific matters such as specific components and limited embodiments and drawings, but these are only provided to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , various modifications and variations are possible from these descriptions by those of ordinary skill in the art to which the present invention pertains.

Therefore, the spirit of the present invention should not be limited to the described embodiments, and not only the claims to be described later, but also all those with equivalent or equivalent modifications to the claims will be said to belong to the scope of the spirit of the present invention. .

110: Computing device capable of processing data arithmetic based on code
111: generative matrix storage unit
112: encoding unit
113: data operation unit
114: decoding unit

Claims (12)

In a computing device capable of processing data arithmetic based on a sign,
A generator matrix in which a generator matrix of a preset size of kxn (k and n are natural numbers greater than or equal to 2 and n is a number greater than k) with error correction capability for t (t is a natural number) bit code is stored. part;
When a first operation command for performing a logical OR on the first message and the second message generated by the computing device is applied, encoding is performed on the first message based on the generation matrix, so that a first codeword ( an encoding unit generating a second codeword by generating a codeword) and encoding the second message based on the generation matrix;
When the first codeword and the second codeword are generated, the first operation codeword is generated by ORing the first codeword and the second codeword, and then the first operation codeword is transmitted to the decoding unit. arithmetic unit; and
By receiving the first operation codeword from the data operation unit and performing error correction-based decoding on the first operation codeword based on the generation matrix, the first message and the second message are logically summed the decoding unit to restore the message
Computing device capable of processing a sign-based data operation comprising a. According to claim 1,
the decoding unit
The first message and the second message are ORed by generating a parity check matrix corresponding to the generation matrix, and performing syndrome decoding on the first operation codeword based on the parity check matrix. A computing device capable of performing a sign-based data operation process for restoring the sum message. According to claim 1,
the encoding unit
When a second operation command to perform convolution on the third message and the fourth message generated by the computing device is applied, encoding the third message based on the generation matrix is performed, generating a fourth codeword by generating 3 codewords and performing encoding on the fourth message based on the generation matrix;
The data operation unit
When the third codeword and the fourth codeword are generated, a Hadamard product of the third codeword and the fourth codeword is performed to generate a second operation codeword, and the second operation codeword is generated. Convolutional codeword of the second operation codeword by multiplying the bootstrap matrix in advance with respect to - The convolutional codeword encodes the result of convolution of the third message and the fourth message based on the generation matrix means the performed codeword - after converting to , the convolutional codeword is transferred to the decoding unit,
the decoding unit
Convolution of the third message and the fourth message by receiving the convolution codeword from the data operation unit and performing error correction-based decoding on the convolution codeword based on the generation matrix A computing device capable of processing sign-based data operations to restore messages. 4. The method of claim 3,
The generating matrix is
It is a generation matrix composed of first-order Reed-Muller (RM) codes,
The bootstrap matrix is
Secondary RM code applied to the second operation codeword - Primary applied to the third codeword and the fourth codeword respectively as the third codeword and the fourth codeword are multiplied by Hadamard RM code is converted from the second operation codeword to the secondary RM code - a message that is converted to the primary RM code and at the same time multiplied by the generation matrix according to the converted primary RM code in the second operation codeword A computing device capable of performing sign-based data arithmetic processing, which is a matrix in which a value is set equal to a result of convolution of the third message and the fourth message. 4. The method of claim 3,
the decoding unit
The convolution message in which the third message and the fourth message are convolved by generating a parity check matrix corresponding to the generation matrix and performing syndrome decoding on the convolution codeword based on the parity check matrix A computing device capable of processing sign-based data operations to restore In the operating method of a computing device capable of processing data arithmetic based on a sign,
A generator matrix in which a generator matrix of a preset size of kxn (k and n are natural numbers greater than or equal to 2 and n is a number greater than k) with error correction capability for t (t is a natural number) bit code is stored. maintaining wealth;
When a first arithmetic command for performing an OR on the first message and the second message generated by the computing device is applied, the encoding unit performs encoding on the first message based on the generation matrix, thereby generating a first code generating a second codeword by generating a codeword and performing encoding on the second message based on the generation matrix;
When the first codeword and the second codeword are generated, a data operation unit generates a first operation codeword by ORing the first codeword and the second codeword, and then transmits the first operation codeword to the decoding unit. delivering; and
The decoding unit receives the first operation codeword from the data operation unit, and performs error correction-based decoding on the first operation codeword based on the generation matrix, so that the first message and the second message are Restoring the disjointed sum message
A method of operating a computing device capable of performing sign-based data arithmetic processing comprising a. 7. The method of claim 6,
The step of restoring the sum message is
The decoding unit generates a parity check matrix corresponding to the generation matrix, and performs syndrome decoding on the first operation codeword based on the parity check matrix, whereby the first message and the second message A method of operating a computing device capable of performing a sign-based data operation processing to restore the sum message in which A is logically ORed. 7. The method of claim 6,
When a second operation command for performing convolution on the third message and the fourth message generated by the computing device is applied, the encoding unit encodes the third message based on the generation matrix generating a third codeword by doing so, and generating a fourth codeword by performing encoding on the fourth message based on the generation matrix;
When the third codeword and the fourth codeword are generated, the data operation unit generates a second operation codeword by performing Hadamard product of the third codeword and the fourth codeword, and the second operation codeword is generated. The second operation codeword is obtained by multiplying two operation codewords by a preset bootstrap matrix. The convolution codeword generates the generation matrix for the result of convolution of the third message and the fourth message. means a codeword on which encoding has been performed based on - converting the convolutional codeword into , and transferring the convolutional codeword to the decoding unit; and
The decoding unit receives the convolution codeword from the data operation unit, and performs error correction-based decoding on the convolution codeword based on the generation matrix, so that the third message and the fourth message are convolutional. Steps to restore the convolutional message
A method of operating a computing device capable of performing sign-based data arithmetic processing further comprising a. 9. The method of claim 8,
The generating matrix is
It is a generation matrix composed of first-order Reed-Muller (RM) codes,
The bootstrap matrix is
Secondary RM code applied to the second operation codeword - Primary applied to the third codeword and the fourth codeword respectively as the third codeword and the fourth codeword are multiplied by Hadamard RM code is converted from the second operation codeword to the secondary RM code - a message that is converted to the primary RM code and at the same time multiplied by the generation matrix according to the converted primary RM code in the second operation codeword A method of operating a computing device capable of performing sign-based data arithmetic processing, which is a matrix whose value is set to be equal to a result of convolution of the third message and the fourth message. 9. The method of claim 8,
The step of restoring the convolutional message is
The decoding unit generates a parity check matrix corresponding to the generation matrix, and performs syndrome decoding on the convolution codeword based on the parity check matrix, whereby the third message and the fourth message are convolutioned. A method of operating a computing device capable of processing a sign-based data operation to restore a convolutional message. A computer-readable recording medium recording a computer program for executing the method of any one of claims 6 to 10 through combination with a computer. A computer program stored in a storage medium for executing the method of any one of claims 6 to 10 through combination with a computer.

Images