TWI551059B - MULTI-CODE CHIEN'S SEARCH CIRCUIT FOR BCH CODES WITH VARIOUS VALUES OF M IN GF(2m) - Google Patents

Description

Multi-mode Qin search circuit for BCH codes with different m values in GF(2 m )

The present invention relates to a multimode Qin search circuit, and more particularly to a multimode Qin search circuit for BCH codes having m values in different GF(2 ^m ).

The Bose-Chaudhuri-Hocquenghem (BCH) code is a very common error correction code used in storage and communication equipment. The BCH code can detect and correct random errors due to noise and defects in the storage device channel. The encoding of the BCH code is often implemented by a combination of a linear feedback shift register and some logic integrated circuits. For the decoding of BCH codes, it is much more complicated than coding. Please refer to Fig. 1. The decoding process can be explained as follows: After receiving a codeword (S01), the codeword is calculated according to a specific polynomial? Value (S02). Then, according to the symptom value, an error position polynomial can be found (S03). Next, by calculating the root of the error position polynomial, an error position number can be obtained (S04). Finally, the error code word is corrected to obtain the correct signal data (S05).

Traditionally, the Peterson-Gorenstein-Zierler (PGZ) algorithm or the Berlekamp-Massey (BM) algorithm can be used to find the above error location polynomial. Because the complexity of the PGZ algorithm is higher than that of the BM algorithm and the BM algorithm can achieve faster decoding speed, the BM algorithm is more popular in hardware implementation. However, due to the need to use multiplicative inverse elements in the BM algorithm, this greatly increases the complexity of the circuit hardware.

According to the error position polynomial λ(x)=λ ₀ +λ ₁ x+...+λ _t x ^t , the root energy of λ(x) can be 1,α,α ² ,...,α ^n-1 (n =2 ^m -1) is simply found in place of λ(x). Since α ⁿ =1 and α ^-1 = α ^n-1 , if α ¹ is an incorrect position number, α ^n-1 is another wrong position number. Traditionally, this substitution process can be iteratively operated by Qin-style search, but this sequential circuit can only find one wrong position number per clock. In order to improve the throughput in the BCH decoding step in each clock, Qin search can utilize the parallel computing architecture. As shown in Figure 2, p error locations are detected in one clock.

Let λ ₁ x+...+λ _t x ^{t in the} error position polynomial λ(x) be Y(x), and if α ⁱ is one root of λ(x), Y(α ⁱ )=1. Under the structure of p parallel operations, the relation of Y(α ^up+i ) can be rewritten as follows: Due to the parallelization, the number of iteration operations becomes [n/p]. Where u represents the number of iterations of the current iteration, and ω _j represents the updated value for each iteration of the operation, 1 i p and 1 j t. When the order of the Galois field is 2 ^m , the Qin search parallel computing architecture in Fig. 2 contains pt finite field multipliers, p t input m finite field adders, t m bits. Register, and t m-bit multiplexers.

If ω _j α ^ij in the above formula is changed to the following matrix: Where Ω _j and A _ij each represent a binary mxm matrix and a binary 1 xm matrix. ω _j α ^ij represents the operation result from the finite field multiplier of the i-th column and the j-th row in Fig. 2, so Y(α ^up+i ) can be derived as: , please note that Y(α ^up+i ) is a 1 xm matrix. In order to be able to represent all of the outputs in Figure 2, a matrix must be generated, one at a time representing all outputs in the u-th iteration, Y(u). Therefore, Y(u) can be expressed in the following matrix: Where Ω(u) is a binary matrix of 1 x mt, representing the intermediate calculated value in the u-th iteration operation, A _y is fixed, independent of the iterative operation. It is worth noting that the Qin search calculations in p parallel computing architectures are formulated in a single matrix of Ω(u) and A _y multiplication.

Further, the updated Ω(u) value can be used in the next iterative operation, resulting in Ω(u+1)=[ω ₁ α ^p ω ₂ α ^2p ... ω _t α ^tp ] Finally, [Y(u)Ω(u+1)]=Ω(u)[A _y A _Ω ] is obtained. The Qin search circuit that implements the p parallel computing architectures of the previous form is shown in Figure 3. This circuit design was published by Youngjoo Lee et al. in August 2011 and is disclosed in IEEE Transactions on circuits and systems-II: Express Briefs, Vol. 58, No. 8, entitled "Low Complexity Parallel Chien's Search Structure Using Two-Dimensional Optimization". In the paper. The output error value vector can be obtained by XORing each output of Y(u) with λ ₀ .

The circuit design described above can only be used for fixed code rates and code lengths in the same GF (2 ^m ). For some applications, it is not applicable when you need to use m values with different code rates, code lengths or even different GF(2 ^m ). Therefore, a new Qin-style search circuit is needed to meet the above requirements.

As described above, in the circuit design of the encoding of the existing BCH code, only the fixed code rate and code length in the same GF (2 ^m ) can be used. Therefore, a new Qin search circuit is needed, which can be applied to m values of different code rates, code lengths and even different GF(2 ^m ).

Therefore, in accordance with an aspect of the present invention, a multi-mode Qin search circuit for BCH codes having m values in different GF(2 ^m ) is proposed. The circuit comprises: a combination matrix unit for providing a plurality of Qin search matrices, receiving a plurality of input values, multiplying the input values by one or all elements of one or more Qin search matrices to obtain a first An operation value and a second operation value, outputting the first operation value, and outputting the second operation values in one of the plurality of line groups according to different input value attributes; the plurality of first multiplexers, each first The multiplexer is connected to a line in each line group and receives a specific second operation value for selecting a second operation value from the corresponding line group according to different input value attributes, and outputting the second operation value a plurality of registers, each of the registers being connected to a specific first multiplexer for receiving the second operation value, and outputting the second operation value in the next clock; and a plurality of a second multiplexer, each of the second multiplexers being connected to a specific register, receiving a specific coefficient value of a polynomial non-linear number of the error location and a second operational value from the temporary register for use in a Qin The first iteration of the search operation In operation, this unit outputs the specific binding matrix coefficient value as the input value, and to the remaining iterative computation formula Qin search operation, a binding matrix to the second computing unit to output the value as an input value. The specific coefficient value for the error position polynomial of the same Qin search matrix and the second operation value calculated for the specific coefficient value have the same input value attribute.

The above-mentioned Qin search matrix has the following form: [A _y A _Ω ], wherein , ,and Where p is the number of parallel operations of the Qin search circuit; t is the error correction capability of the corresponding BCH code; α ₀ , α ₁ , ... and α _m-1 are a standard base in GF(2 ^m ) ;1 i p;1 j t.

According to the concept of the case, different Qin search matrices correspond to different m and / or t values. Each Qin search matrix is divided into a plurality of parts of the number of error correction capabilities corresponding to the corresponding BCH code in the column direction, and each part contains sub-matrices corresponding to Ay and AΩ. The parts of a Qin search matrix are sequentially aligned with the parts of another Qin search matrix in a manner. The mode is one-side alignment, center-aligned, or offset from the side by a certain amount. The positions in the combination matrix unit that are not covered by the parts of the Qin search matrix are complemented by zeros.

Moreover, the Qin search matrices are separated by 0, and some elements between the two Qin search matrices have a common subexpression, and one Qin search matrix uses a common subexpression of another Qin search matrix. The first calculated value is further added to the constant term value of the error position polynomial. If the added value is 0, the corresponding element α ^up+i , 1 i p, which is a root of the polynomial of the error location.

Therefore, according to the above, by designing a Qin search circuit with several Qin search matrices, and using peripheral components, the use of m values for different code rates, code lengths, and different GF(2 ^m ) can be achieved. purpose. At the same time, the hardware complexity of the Qin-style search circuit can also be reduced.

100‧‧‧ Qin search circuit

120‧‧‧Combined matrix unit

140‧‧‧First multiplexer

160‧‧‧ register

180‧‧‧Second multiplexer

200‧‧‧ Qin search circuit

220‧‧‧Combined matrix unit

240‧‧‧First multiplexer

260‧‧‧ register

280‧‧‧Second multiplexer

A‧‧‧Qin search matrix A

B‧‧‧Qin search matrix B

C‧‧‧Qin search matrix C

A1-A7‧‧‧Parts of the Qin search matrix A

B1-B4‧‧‧Parts of the Qin search matrix B

C1-C5‧‧‧Parts of the Qin search matrix C

Figure 1 is a flow chart of a conventional decoding method for a BCH code.

Figure 2 shows the circuit implemented by a traditional Qin search.

Figure 3 shows the circuit implemented by another traditional Qin search.

Figure 4 is an embodiment of a multi-mode Qin search circuit in accordance with the present invention.

Fig. 5 is a diagram showing a combination matrix unit configuration in this embodiment.

Figure 6 illustrates another combined matrix unit configuration in this embodiment.

Figure 7 is a further embodiment of a multi-mode Qin search circuit in accordance with the present invention.

Fig. 8 is a diagram showing a combination matrix unit configuration in this embodiment.

The invention will be more specifically described by reference to the following examples.

Please refer to Figures 4 through 6, to illustrate an embodiment in accordance with the present invention. A multi-mode Qin search circuit 100 can be used for BCH codes having m values in different GF(2 ^m ). In the present embodiment, two different m values, m ₁ and m ₂ , (m ₁ &gt _; m ₂ ) are used for explanation. The BCH code obtained based on these two m values has different code lengths and code rates (because m ₁ > _m2 , the relative code length is 2 ^m1 >2 ^m2 , but the code rate relationship is not necessarily), but each has t ₁ and t ₂ error correction capabilities. In the present embodiment, t ₁ &gt _; t ₂ is set. As shown in FIG. 4, the Qin type search circuit 100 includes one combining matrix unit 120, t first multiplexers 140, t register 160, and t second multiplexers 180. Each of the following describes the role of these elements.

The combining matrix unit 120 is the most important core component and can be used to provide more than two Qin search matrices. In this embodiment, the combining matrix unit 120 includes two Qin search matrices, a Qin search matrix A and a Qin search matrix B. Qin type search matrix A is used to find the solution of the error position polynomial a, and the coefficient of the error position polynomial a is obtained by receiving the BCH code C ₁ with t ₁ error correction ability, finding the symptom value, and Licam-Messi algorithm to find out. Similarly, the Qin search matrix B is used to find the solution of the error position polynomial b, and the coefficient of the error position polynomial b is obtained by receiving the BCH code C ₂ with t ₂ error correction capabilities, and finding the syndrome value, and Find out by the Berkamp-Messi algorithm. The structure of Qin-style search matrix A and Qin-style search matrix B will be introduced later. It should be noted, however, that the bonding matrix unit 120 can be any electronic component, such as a read only memory (ROM) array, to memorize the elements (0 or 1) of the Qin search matrix.

The combining matrix unit 120 can receive a plurality of input values. As shown in FIG. 4, the input values are derived from the second multiplexer 180. The combining matrix unit 120 multiplies the input values by some or all of the elements in one or more Qin search matrices to obtain a first operational value and a second operational value. In this embodiment, although the Qin search matrix A is used for BCH coding C ₁ , the Qin search matrix B is mainly used for BCH coding C ₂ , but since the Qin search matrix A and the Qin search matrix B have common elements. The expression, Qin type search matrix B can use the Qin type to search for some elements of matrix A.

The first operational value referred to herein, that is, Y(α ^up+i ) mentioned in the prior art, the result output after each iterative operation (u is the number of iterations, p is the parallel matrix structure of the combining matrix unit 120) The number of α ^up+i is one of the possible solutions of Qin style search, 1 i p), by determining whether the addition of the constant term value λ _{0 of the} error position polynomial is 0, determines whether the corresponding element α ^up+i is one root of the error position polynomial. The second operation value is an intermediate value in the recursive operation process, and is input to the combination matrix unit 120 for calculation. The combining matrix unit 120 may output the first operational value and output the second operational values in one of the two line groups according to different input value attributes. The specific coefficient value for the error location polynomial of the same Qin search matrix and the second operation value calculated for the specific coefficient value have the same input value attribute. That is, when the coefficient values for the same Qin search matrix are input through the second multiplexer 180 in the clock of the first recursive operation, the second operation value calculated by the coefficient values is included. , all have the same input value attribute. For the convenience of operation of the first multiplexer 140, the output of each second operational value has a specific line for the second operational value of the different input value attributes, that is, the first multiplexer in FIG. The two upper and lower two lines connected to the 140 are input to the second operation value ΩA ₁ (u) and ΩA ₂ (u)... ΩA _t (u) calculated by the Qin search matrix A. For the group, the path for inputting the second operational value ΩB ₁ (u), ΩB ₂ (u), ΩB _t (u) calculated by the Qin search matrix B is another set of circuit groups. It is worth noting that the number of line groups is not limited to two, and any number is feasible, but it is equal to or larger than the number of Qin search matrices.

In this embodiment, the number of the first multiplexers 140 is t, and t is greater than or equal to the larger of t ₁ and t ₂ so that the maximum correction capability of the BCH code can be realized, and t is also the error position of the operation. The highest order of the polynomial. As previously mentioned, each first multiplexer 140 is coupled to a line in each line group and receives a particular second operational value from the combining matrix unit 120. The first multiplexer 140 may select a second operational value from the corresponding line group according to different input value attributes, and output the second operational value to the connected register 160.

Each of the registers 160 is connected to a specific first multiplexer 140, and receives the second calculated value, and outputs the second calculated value to the connected terminal in the next clock. The second multiplexer 180. Due to the operation of the register 160, the Qin search circuit 100 can perform an iterative operation in each clock.

Each second multiplexer 180 is coupled to a particular register 160 to receive a particular coefficient value of a polynomial non-linear term of the error location (ie, λ ₁ , λ ₂ ... λ _{n in} FIG. 4 ) And a second operational value from the register 160. The function of the second multiplexer 180 is to output the specific coefficient value to the combining matrix unit 120 as the input value in the first iterative operation starting from a Qin-type search operation; and in the iterative operation of the remaining Qin-style search operations. The second operational value is output to the combining matrix unit 120 as an input value.

The structure of the Qin search matrix A and the Qin search matrix B is described here. The Qin-style search matrix A has the following form: [A _y A _Ω ], where Where p is the number of parallel operations of the Qin search circuit, t ₁ is the error correction capability of the BCH code C ₁ , α ₀ , α ₁ , ... and For GF ( a standard substrate, 1 i p,1 j t ₁ . The Qin-style search matrix B has the following form: [B _y B _Ω ], where , where t ₂ is the error correction capability of BCH code C ₂ , α ₀ , α ₁ , ... and For GF ( a standard substrate, 1 j t ₁ .

Since t _{1 is} greater than t ₂ , the matrices A _y and A _{Ω are} necessarily greater than B _y and B _Ω . Since the Qin type search matrix B needs to share some elements of the Qin search matrix A, in order to facilitate the operation, the arrangement has a special design. See Figure 5. Each Qin search matrix is divided into several parts of the number of error correction capabilities (t ₁ and t ₂ ) of the corresponding BCH code in the column direction. Each part of the Qin search matrix A contains a corresponding A _y and A _{Ω .} For the sub-matrix matrix, each part of the Qin-style search matrix B contains sub-matrices corresponding to B _y and B _Ω . Here further let t ₁ =7, t ₂ =4, then the Qin search matrix A is divided into 7 parts, such as A1, A2...A7, and the Qin search matrix B is divided into B1, B2...B4, etc. 4 parts. Each part of the Qin-style search matrix A contains m ₁ bit elements in terms of length, and each part of the Qin-style search matrix B contains m ₂ bit elements.

It should be noted that this figure, together with the sorting of all the patterns describing the Qin search matrix, does not necessarily have to be drawn according to the actual scale, only for the purpose of its application.

The parts of the Qin-style search matrix B are sequentially aligned with the parts of the Qin-style search matrix A in a specific manner. As shown in Fig. 5, the upper side of each part of the Qin search matrix B is aligned with the upper side of a portion corresponding to the Qin search matrix A. For example, the upper side of A1 is aligned with the upper side of B1, the upper side of A2 is aligned with the upper side of B2, and the like. It should be noted that since the Qin search matrix B is smaller than the Qin search matrix A, the parts of the Qin search matrix B are basically separated, but the parts of the Qin search matrix A are also substantially connected together (dotted line To mark each part). The bonding matrix unit 120 is based on circuit design considerations. The positions not covered by the various parts of the Qin search matrix are supplemented by 0. However, the Qin search matrix A and the Qin search matrix B can also be separated by 0 for easy identification, as shown in FIG. The specific manner described above may be center aligned or offset from the side by a certain amount, except for one side alignment. It is worth noting that there are many common methods that can be found between the Qin-style search matrix A and the Qin-style search matrix B. Since it is not within the scope of the invention, it is omitted.

It can be seen from the above that in the Qin type search circuit 100, different Qin search matrices correspond to different m values (m ₁ and m ₂ ) or t values (t ₁ and t ₂ ). But in practice, it can be one of them. In addition, the combination matrix unit 120 can complete the design by first establishing A _y and A _Ω ; secondly establishing B _y and B _Ω ; finally finding a common sub-expression to establish a binding matrix.

In accordance with the spirit of the present invention, the Qin search circuit may also include three or more sets of lines. In this case, the arrangement of the Qin search matrix will be different. Please see the description of another embodiment below.

Referring to Figure 7, another multi-mode Qin search circuit 200 can be used for BCH codes having m values in different GF(2 ^m ). In the present embodiment, a third different m value other than the foregoing embodiment, m ₃ , (m ₁ > m ₃ > m ₂ ) is used, since m ₁ > m ₃ &gt _; m ₂ , the relative code length 2 ^m1 >2 ^m3 >2 ^m2 . The BCH codes corresponding to m ₁ , m ₂ and m ₃ each have t ₁ , t ₂ and t ₃ error correction capabilities. In the present embodiment, t ₁ > t ₃ > t _{2 is set} such that t ₁ = 7, t ₂ = 4, and t ₃ = 5. The Qin-type search circuit 200 also includes a combination matrix unit 220, t first multiplexers 240, t registers 260, and t second multiplexers 280, each of which describes the differences between the components. Corresponds to the function of the elements of the previous embodiments.

The combining matrix unit 220 is the most important core component, and includes three Qin search matrices, a Qin search matrix A, a Qin search matrix B, and a Qin search matrix C. The Qin search matrix A and the Qin search matrix B are the same as the previous embodiment. The Qin search matrix C is used to find the solution of the error location polynomial c, and the error location polynomial c is obtained by receiving t ₃ coefficients. The error correction capability BCH encodes C ₃ , finds its syndrome value, and finds it by the Berkamp-Messi algorithm. The combining matrix unit 220 can receive a plurality of input values from the second multiplexer 280 and multiply the input values by one or all of the elements in the one or more Qin search matrices to obtain the first calculated value and the first The second operation value. In this embodiment, the Qin search matrix A and the Qin search matrix B have common subexpressions, and the Qin search matrix A and the Qin search matrix C also have a common subexpression. Therefore, both the Qin search matrix B and the Qin search matrix C can use some elements of the Qin search matrix A, but there is no common subexpression between the two.

The first operational value and the second operational value referred to herein are similar to those defined in the previous embodiments, which are mentioned in the prior art. The difference is that the bonding matrix unit 220 replaces the bonding matrix unit 120. The combining matrix unit 220 may output the first operational value and output the second operational values in one of the three line groups according to different input value attributes. When the coefficient values for the same Qin search matrix are input through the second multiplexer 280 in the clock of the first recursive operation, the second operation value calculated by the coefficient values is then included. Has the same input value attribute. For the convenience of operation of the first multiplexer 240, the output of each second operational value has a specific line for the second operational value of the different input value attributes, that is, the first multiplexer in FIG. The upper, middle and lower three lines connected to the 140 are input to the second operational value ΩA ₁ (u) and ΩA ₂ (u)... ΩA _t (u) calculated by the Qin search matrix A. Belong to the same group of lines, input the path of the second operation value ΩB ₁ (u), ΩB ₂ (u)... ΩB _t (u) calculated by the Qin search matrix B is another group of lines. The path for inputting the second operational value ΩC ₁ (u), ΩC ₂ (u)... ΩC _t (u) calculated by the Qin search matrix C is another set of circuit groups.

In this embodiment, the number of the first multiplexers 240 is t, and t is greater than or equal to the larger of t ₁ , t _{2 ,} and t ₃ , so that the maximum correction capability of the BCH code can be realized, and t is also an operation. The highest order of the error location polynomial. As previously mentioned, each first multiplexer 240 is coupled to a line in each line group and receives a particular second operational value from the combining matrix unit 220. The first multiplexer 240 may select a second operational value from the corresponding line group according to different input value attributes, and output the second operational value to the connected register 260.

Each of the registers 260 is connected to a specific first multiplexer 240, and receives the second calculated value, and outputs the second calculated value to the connected terminal in the next clock. The second multiplexer 280. Each second multiplexer 280 is coupled to a particular register 260 for receiving a particular coefficient value of a polynomial non-linear term of an error location (ie, λ ₁ , λ ₂ ... λ _{n in} FIG. 7) And a second operational value from the register 260. The function of the second multiplexer 280 is to output the specific coefficient value to the combining matrix unit 220 as the input value in the first iterative operation starting from a Qin-style search operation. As described herein; and in the iterative operation of the remaining Qin-style search operations, the second operational value is output to the combining matrix unit 220 as an input value.

The structure of the Qin search matrix A and the Qin search matrix B in this embodiment is the same as the previous embodiment. The Qin-style search matrix C has the following form: [C _y C _Ω ], where Where p is the number of parallel operations of the Qin search circuit, t ₃ is the error correction capability of the BCH code C ₃ , α ₀ , α ₁ , ... and For GF ( a standard substrate, 1 i p,1 j t ₃ .

Since t _{1 is} greater than t ₃ and t ₃ is greater than t ₂ , the matrices A _y and A _{Ω are} necessarily greater than C _y and C _Ω , and C _y and C _{Ω are} greater than B _y and B _Ω . See Figure 8 for the arrangement of the Qin search matrix. Each Qin search matrix is divided into several parts of the error correction capability (t ₁ , t ₂ and t ₃ ) of the corresponding BCH code in the column direction. Each part of the Qin search matrix A contains a corresponding A _{y .} With the sub-matrix of A _Ω , each part of the Qin search matrix B contains sub-matrices corresponding to B _y and B _Ω, and each part of the Qin search matrix C contains sub-matrices corresponding to C _y and C _Ω . As described above, t ₁ =7, t ₂ =4, and t ₃ =5, then the Qin search matrix A is divided into seven parts, A1, A2, ..., A7, and the Qin search matrix B is divided into B1, B2. In the four parts of .B4, the Qin-style search matrix C is divided into five parts: C1, C2...C5. In terms of length, the Qin search matrix A and the Qin search matrix B are not repeated as described above, and the Qin search matrix C contains m ₃ bit elements.

The parts of the Qin-style search matrix C are the same as the parts of the Qin-style search matrix B, and the parts of the Qin-style search matrix A are sequentially aligned in a specific manner. As shown in Fig. 8, the upper side of each part of the Qin search matrix C is aligned with the upper side of the upper edge of the portion corresponding to the Qin type search matrix A. For example, the upper side of A1 is aligned with the upper side of C1, the upper side of A5 is aligned with the upper side of C5, and the like. Since the Qin search matrix A is smaller than the Qin search matrix A, the parts of the Qin search matrix C are basically separated, and the positions of the combination matrix unit 220 that are not covered by the parts of the Qin search matrix are 0. Make up.

In this embodiment, some elements between the two-Qin search matrix (Qin search matrix A and Qin search matrix B, or Qin search matrix A and Qin search matrix C) have common sub-expressions, one Qin The search matrix can use a common subexpression of another Qin search matrix. However, the Qin-style search matrix B and the Qin-style search matrix C have no common sub-expressions. In practice, the Qin-style search matrix B and the Qin-style search matrix C can also have a common sub-expression. Even, there is a common subexpression between the three Qin search matrices, but all three have no common subexpression.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the invention, and any one of ordinary skill in the art can make some modifications and refinements without departing from the spirit and scope of the invention. The scope of the invention is defined by the scope of the appended claims.

100‧‧‧ Qin search circuit

120‧‧‧Combined matrix unit

140‧‧‧First multiplexer

160‧‧‧ register

180‧‧‧Second multiplexer

Claims (10)

A multi-mode Qin search circuit for a BCH code having different m values in GF(2 ^m ), comprising: a combining matrix unit for providing a plurality of Qin search matrices, receiving a plurality of input values, inputting the inputs Multiplying a value with one or more elements of one or more Qin search matrices to obtain a first operational value and a second operational value, outputting the first operational value, and according to different input value attributes, in a plurality of line groups Outputting the second calculated values; a plurality of first multiplexers, each of the first multiplexers being connected to a line in each line group and receiving a specific second operation value for different input values Attribute, selecting a second operation value from the corresponding line group, and outputting the second operation value; a plurality of registers, each register being connected to a specific first multiplexer for receiving the a second operation value, and outputting the second operation value in a next clock; and a plurality of second multiplexers, each of the second multiplexers being connected to a specific register, receiving a polynomial of a wrong position polynomial a specific coefficient value and the number from the register The operation value is used to output the specific coefficient value to the binding matrix unit as the input value in the first iterative operation starting from a Qin-type search operation, and in the iterative operation of the remaining Qin-style search operations, to the combination The matrix unit outputs the second operational value as the input value, wherein the specific coefficient value for the error location polynomial of the same Qin search matrix and the second operational value calculated for the specific coefficient value have the same input value property. For example, the Qin type search circuit described in claim 1 wherein the Qin search matrix has the following form: [A _y A _Ω ], wherein Where p is the number of parallel operations of the Qin search circuit; t is the error correction capability of the corresponding BCH code; α ₀ , α ₁ , ... and α _m-1 are a standard base in GF(2 ^m ) ;1 i p;1 j t. For example, the Qin type search circuit described in claim 2, wherein different Qin search matrices correspond to different m and/or t values. For example, in the Qin type search circuit described in claim 2, each of the Qin search matrices is divided into a plurality of parts of the number of error correction capabilities corresponding to the corresponding BCH code in the column direction, and each part contains a corresponding A _{y Submatrix} equal to A _Ω . For example, in the Qin type search circuit described in claim 4, each part of a Qin search matrix is sequentially aligned with each part of another Qin search matrix in a manner. The Qin type search circuit as described in claim 5, wherein the mode is one side aligned, center aligned or offset from the side by a certain amount. The Qin-type search circuit of claim 5, wherein the position of the combination matrix unit that is not covered by the parts of the Qin search matrix is complemented by zero. For example, the Qin type search circuit described in claim 5, wherein the Qin search matrices are separated by 0. For example, in the Qin type search circuit described in claim 1, wherein some elements between the two Qin search matrices have a common sub-expression, and one Qin search matrix uses a common sub-expression of another Qin search matrix. The Qin type search circuit of claim 1, wherein the first operation value is further added to the constant term value of the error position polynomial, and if the addition value is 0, the corresponding element α ^up+i , 1 i p is a root of the polynomial of the error location, where α is the number of error positions, u is the number of iterations, and α and u are positive integers, respectively.