A low-complexity hardware system and application method based on CORDIC
Through a low-complexity hardware system based on CORDIC, the pipeline architecture and parallel result processing unit are used to dynamically support complex root calculations of 2 to 10th power, solving the complex N power root calculation problems, achieving high efficiency, Accurate computing speed improvement and resource conservation.
Patent Information
- Application Number
- CN202111135783.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-09-27
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2041-09-27
AI Technical Summary
In the prior art, complex N-power root calculations are time-consuming and inefficient in software implementation, hardware implementation is complex and the number of results is uncertain. Especially when high-order root calculations, the circuit complexity increases sharply and cannot meet the wide range of application needs.
Using a low-complexity hardware system based on CORDIC, the pipeline architecture and parallel result processing unit uses the similarity of the N roots of complex N power to solve the process, dynamically supports the calculation of complex 2 to 10th power, reducing the calculation complexity and resource consumption to achieve high efficiency and high accuracy.
It realizes fast and high-precision calculations of complex numbers 2 to 10th power roots, expands the calculation convergence range, improves throughput and flexibility, increases the circuit complexity by only 0.157%, increases the calculation speed by 379 times, supports a wide range and error Less than 10-6.
Smart Images

Figure CN113778379B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of complex number Nth root operation, and more specifically, to a low-complexity hardware system and application method based on CORDIC. Background Art
[0002] Complex number operations are the core part of circuit calculations and are widely used in communication systems and signal processing fields for real-time data representation and system modeling. Nth root operations are an important part of complex function theory. In polynomial calculations, matrix calculations, trigonometric functions, etc., timely introduction of Nth root operations can simplify the calculation process. However, complex Nth root operations are very complex due to the uncertainty of the number of roots and the complexity of complex operations. Most Nth root operation research focuses on real numbers, or usually implements complex Nth root operations by software. However, software implementation often uses mixed operations of various algorithms rather than dedicated algorithms to ensure accurate and reliable calculations. There is often redundancy in the process and it performs poorly in real-time work.
[0003] Another method is to accelerate the Nth root operation through application-specific integrated circuit (ASIC) hardware to achieve high computing performance. However, there are only a few works related to the hardware implementation of complex square roots. As high-order roots are widely used in atmospheric models, radiation and other fields, square root operations alone can no longer meet application requirements. In specific hardware implementations, the resources consumed by the constant associated with N are proportional to the range of N. In practical applications, the most commonly used N value is an integer from 2 to 10.
[0004] A coordinate rotation digital computer (CORDIC) can effectively calculate transcendental functions such as trigonometric functions, exponential functions, and logarithmic functions through simple shift and addition operations, can achieve a high calculation speed, and has a good trade-off between accuracy and area, and achieves low cost. To this end, the present invention proposes a low-complexity hardware solution based on CORDIC for calculating the 2nd to 10th roots of complex numbers, achieving high calculation efficiency while reducing hardware implementation complexity.
[0005] Because the number of results is uncertain, Nth root calculation has always been a challenging topic, so most hardware implementations focus on real number Nth roots or just quadratic root calculations. Chinese patent application number CN202011357034.8, published on March 12, 2021, discloses a method for calculating the Nth square root of a complex number based on the CORDIC method. This method is the result of the preliminary work of the applicant of the present invention. Although it can realize the calculation of Nth roots of arbitrary orders, it can only calculate the Nth root of each input complex number serially according to the value of N, and only calculate one of the N results each time. However, the calculation of the complex number Nth root has N results. When the value of N is large, the complexity of the circuit will increase rapidly with the increase of the value of N. Therefore, the calculation efficiency and flexibility of this method are insufficient. Summary of the invention
[0006] 1. Technical issues to be solved
[0007] In view of the problems in the prior art that the software implementation of complex Nth root operations is time-consuming and inefficient, and the hardware implementation of complex Nth root operations is complex and the number of results is uncertain, the present invention provides a low-complexity hardware system and application method based on CORDIC, which adopts a pipeline architecture and uses the similarity of the N roots of the complex Nth root to reduce the calculation cost. It has the characteristics of wide support range, high efficiency, high precision and low hardware complexity.
[0008] 2. Technical solution
[0009] The purpose of the present invention is achieved through the following technical solutions.
[0010] A low-complexity hardware system based on CORDIC, comprising a circular vector mode CV-CORIDC module, a circular rotation mode CR-CORDIC module, a first linear vector mode LV1-CORDIC module, a second linear vector mode LV2-CORDIC module, a hyperbolic vector mode HV-CORDIC module, a hyperbolic rotation mode HR-CORDIC module and a result processing unit;
[0011] The input data of the system is input to the input end of the CV-CORIDC module, the output end of the CV-CORIDC module is connected to the input end of the HV-CORDIC module and the input end of the LV2-CORDIC module, the output end of the HV-CORDIC module is connected to the input end of the LV1-CORDIC module, the output end of the LV1-CORDIC module is connected to the input end of the HR-CORDIC module, the output end of the HR-CORDIC module and the output end of the LV2-CORDIC module are both connected to the input end of the CR-CORDIC module, the output end of the CR-CORDIC module is connected to the input end of the result processing unit, and the result processing unit outputs the calculation result.
[0012] The present invention uses an efficient parallel result processing unit, utilizes the similarity of the solution process of the Nth root of a complex number to reduce the calculation complexity, and adopts a pipeline architecture, which can dynamically support the calculation of the 2nd to 10th roots of complex numbers, saves storage resources, and overcomes the problems of high calculation complexity and long calculation time of the Nth root.
[0013] Furthermore, the system uses a pipeline architecture.
[0014] Furthermore, the result processing unit includes a plurality of parallel result units.
[0015] The present invention uses three CORDIC modules, namely circular, linear and hyperbolic, and a parallel result processing unit to construct an efficient hardware implementation method; the present invention adopts a pipeline architecture to achieve high-speed full-pipeline calculation.
[0016] A method for calculating the 2 to 10th power root of a complex number, using the CORDIC-based low-complexity hardware system; the expression of the complex number z is: z=m+j*n, and the Nth power root of the complex number z is calculated by the exponential form of the complex number, where N is an integer greater than or equal to 2 and less than or equal to 10.
[0017] Specific: in d=0,1,…,N-1;when m≥0, When m<0,n≥0, When m<0,n<0,
[0018] Furthermore, the result processing unit obtains the sum and difference of two angles through shift and addition operations using the trigonometric function formula. The system inputs the real part m and imaginary part n of the complex number z, and the settlement calculation unit calculates The real part of and the imaginary part
[0019] Furthermore, according to N different d values, There are N roots, d = 0, 1, ..., N-1; when the value of d is determined, and is a constant that is calculated in advance and stored in a lookup table.
[0020] Furthermore, the integer N value input in each clock cycle changes dynamically between 2 and 10, and the result processing unit dynamically activates different result units according to the different input integers N, and completes the calculation of N roots in parallel. The present invention can dynamically support the calculation of the 2nd to 10th roots of complex numbers, with up to 10 parallel calculations.
[0021] Furthermore, the present invention innovatively uses an efficient parallel result processing unit, and utilizes the similarity in the process of solving N roots to share the computing resource CORDIC module, thereby reducing the hardware implementation complexity of high-order Nth root. The calculation results of the CORDIC module are output to the parallel result processing unit, thereby reducing the hardware complexity. The problem of high Nth root calculation complexity and long calculation time is overcome.
[0022] Further, each CORDIC module is iterated X times through an X-stage pipeline architecture to complete the result convergence, and the number of iterations includes the positive number of iterations for determining the calculation accuracy and the negative number of iterations for expanding the calculation convergence range, and X is an integer greater than zero. According to the simulation results, the present invention selects 23 iterations with higher accuracy, and performs 23 iterations on each CORDIC module through a 23-stage pipeline architecture to complete the result convergence, the maximum number of negative iterations is -2, and the maximum number of positive iterations is 20, wherein the positive number of iterations determines the accuracy, and the negative iteration is used to expand the calculation convergence range. For the three coordinate systems, the negative expansion method is slightly different:
[0023] (a) For the circular CORDIC algorithm, the sequence given by k = 0, 0, 0, 1, ..., 20 has been verified to be a better sequence than k = -2, -1, 0, 1, ..., 20;
[0024] (b) For the linear CORDIC algorithm, we expand the iteration index set to k = -2, -1, 0, 1, ..., 20;
[0025] (c) For the hyperbolic CORDIC algorithm, we expand the set of iteration indices to k = -2, -1, 0, 1, ..., 20, just like the linear CORDIC algorithm. The difference from the linear CORDIC algorithm is that the negative iteration has an iteration formula independent of the positive iteration operation.
[0026] The present invention adopts a pipeline architecture, which can calculate multiple iterative processes at the same time and improve the throughput. Taking the 10th root operation of a complex number as an example, after an earlier calculation, it can calculate 10 roots in one clock cycle.
[0027] Furthermore, the constant values related to N are calculated and stored in the lookup table, and the corresponding calculated values are obtained according to different integers N between 2 and 10. and The constant values related to N are stored in the lookup table, and these constants will be read accordingly in each iteration to save hardware calculation time, reduce hardware overhead, and achieve flexible calculation.
[0028] 3. Beneficial effects
[0029] Compared with the prior art, the present invention is a low-complexity hardware system based on CORDIC, and discloses an architecture that dynamically supports complex number 2 to 10 root operations, which can effectively improve the input range, throughput and flexibility, and reduce the hardware implementation complexity and resource consumption, which is specifically manifested in:
[0030] (1) The present invention expands the computational convergence range, and the software simulation can support 10 -8 to 10 4 The relative error can reach 10 -6 , supports a wide range and has high calculation accuracy;
[0031] (2) The present invention adopts a pipeline architecture, which can calculate multiple iterative processes at the same time, thereby improving the throughput rate; in the calculation process, the constant values related to N are calculated in advance and stored in the lookup table, and the corresponding constants are flexibly retrieved according to different integers N between 2 and 10, thereby further reducing the hardware overhead and saving storage resources;
[0032] (3) The integer N value input in each clock cycle of the present invention can be dynamically changed between 2 and 10. According to the different input integers N, different result processing units are dynamically activated to calculate N results in parallel, thereby ensuring flexibility and improving calculation efficiency.
[0033] (4) The present invention uses the similarity of the process of solving the N roots of a complex number to share computing resources and reduce the hardware implementation complexity of high-order N roots; it uses the CORDIC characteristics to solve the problem through simple shift and addition operations. Taking the calculation of the 10th root of a complex number as an example, after earlier calculations, the present design can calculate 10 roots in one clock cycle. Compared with calculating only one of the roots of the second root of a complex number, the circuit complexity is only increased by 0.157%; compared with the prior art, the calculation speed of the present invention can be increased by 379 times. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 It is a schematic diagram of the hardware architecture of the present invention;
[0035] Figure 2 It is a schematic diagram of the parallel result processing unit architecture of the present invention;
[0036] Figure 3 23-stage pipeline architecture diagram of the CV-CORDIC module of the present invention;
[0037] Figure 4 It is a simulation relationship diagram of the average relative error, integer N value and positive iteration number P in the present invention. DETAILED DESCRIPTION
[0038] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. However, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features known in the art are not described.
[0039] Example
[0040] The low-complexity hardware system architecture of this embodiment is as follows Figure 1 As shown, the system inputs the real part m and imaginary part n of the complex number z, and passes through three CORDIC algorithms: circular, linear and hyperbolic. As shown in the figure, the system includes a circular vector mode CV-CORIDC module, a circular rotation mode CR-CORDIC module, a first linear vector mode LV1-CORDIC module, a second linear vector mode LV2-CORDIC module, a hyperbolic vector mode HV-CORDIC module and a hyperbolic rotation mode HR-CORDIC module; the calculation data is first input into the CV-CORIDC module, the output data of the CV-CORDIC module is input into the HV-CORDIC module and the LV2-CORDIC module, the output data of the HV-CORDIC module passes through the LV1-CORDIC module and the HR-CORDIC module in turn, and flows to the CR-CORDIC module, the output data of the LV2-CORDIC module also flows to the CR-CORDIC module, and the output data of the CR-CORDIC module is sent to the result processing unit to obtain the result through shift and addition operations. and As shown in the figure, the system adopts a full-pipeline architecture to achieve high-speed full-pipeline computing.
[0041] This embodiment uses three coordinate systems and five CORDIC algorithms to calculate the Nth root of a complex number. Each coordinate system includes two modes: rotation and vector. The generalized CORDIC formula is as follows:
[0042] x k+1 =x k -μd k (2 -k y k )
[0043] y k+1 =y k +d k (2 -k x k )
[0044] z k+1 =z k -d k e k
[0045] Where k represents the current iteration number, d k represents the decision operator, d k The value of d is determined by the CORDIC operating mode. In rotation mode, d k =sign(z k ); in vector mode, d k =sign(x k y k ). In addition, when different coordinate systems are used to describe the CORDIC equations, μ and e k The value of is shown in the following formula:
[0046]
[0047] In the above formula, circlar represents circular coordinate system, linear represents linear coordinate system, and hyperbolic represents hyperbolic coordinate system. In general, k starts from 0, and the k value increases by 1 each iteration. It can also be expanded by negative expansion, that is, k starts from a negative value to expand the calculation range; an exception to the hyperbolic coordinate system is that when k = 4, 13, 40, these fixed special values, the iteration needs to be repeated once to ensure convergence.
[0048] This embodiment performs 23 iterations for each CORDIC operator to complete the result convergence. Theoretically, without considering the accuracy, the number of iterations of this system can be any value. This embodiment takes 23 times as an example, based on the software simulation, it is found that 23 iterations can achieve higher accuracy. The maximum number of positive iterations is selected as 20, at which time the relative error can reach 10 -6 , which is sufficient to meet the requirements of good accuracy; at the same time, the maximum index of negative iteration is -2, which can extend the input range to 10 -8 to 10 4 , which satisfies the Nth root operation of general data. After 23 iterations, the CORDIC equation converges to the value shown in Table 1.
[0049] Table 1 Output convergence values of CORDIC in three coordinate systems and two modes
[0050]
[0051]
[0052] In Table 1, χ and λ are scaling factors used to correct the results. For a circular coordinate system, For a hyperbolic coordinate system, Since the number of iterations is known, these scaling factors are constants and can be pre-computed by software and stored in a lookup table to avoid complex hardware calculations.
[0053] According to Table 1, the Nth root of the complex number z=m+j*n is calculated using the CV-CORDIC module, HV-CORDIC module, LV1-CORDIC module, HR-CORDIC module, LV2-CORDIC module and CR-CORDIC module. The calculation result of the CV-CORDIC module needs to pass through a multi-level buffer unit before the data is sent to the LV2-CORDIC module for calculation to ensure synchronization with the HR-CORDIC module. The connection method of each CORIDC module operator and the input and output values are as follows Figure 1 As shown, the result processing unit performs shift and addition operations to finally output and
[0054] The exponential form of z is ρe j(2dπ+θ) , solve the Nth root of the complex number z through exponential form; then:
[0055] in d=0,1,…,N-1;when m≥0, When m<0,n≥0, When m<0,n<0,
[0056] According to N different d values, There are N roots; when the value of d is determined, and is a constant, which is stored in the lookup table in advance, and the corresponding constant is flexibly taken out according to different integers N between 2 and 10 to save hardware calculation time.
[0057] like Figure 2 As shown, the result processing unit adopts a parallel structure, and the trigonometric function two-angle sum and difference formula is used in the parallel result processing unit to calculate The real part of and the imaginary part
[0058]
[0059]
[0060] There are N roots for high-order Nth roots. As N increases, the complexity of the circuit increases rapidly. Taking N=10 as an example, if 10 computing resources are used to calculate 10 results in parallel, the hardware overhead is 10 times the original; if 10 roots are calculated serially, a lot of computing time will be consumed. In order to reduce the complexity of the high-order Nth root, this embodiment uses the similarity in the process of solving the N roots to share computing resources, and finally connects the parallel result processing units for parallel processing, thereby reducing the hardware implementation complexity of the high-order Nth root.
[0061] The integer N value input in each clock cycle can change dynamically between 2 and 10. The parallel result processing unit dynamically activates different result units in the result processing unit according to the different input integers N to complete the calculation of N roots in parallel, such as Figure 2 As shown, at this time N=3, result1 to result3 units are activated in the result processing unit, and 3 roots are calculated at the same time.
[0062] Figure 3 This is the 23-stage pipeline architecture diagram of the CV-CORDIC module in the present invention. The architectures of other CORDIC module operators are similar. In the first three reverse iterations, k = 0, 0, 0, and the iteration formula at this time is:
[0063] x k+1 =x k -sign(y k )*y k
[0064] y k+1 =y k +sign(y k )*x k
[0065] z k+1 =z k -sign(y k )*tan -1 (1)
[0066] In the next twenty forward iterations, k = 1, 2, ..., 20, and the iteration formula is:
[0067] x k+1 =x k -sign(y k )*y k *2 -k
[0068] y k+1 =y k +sign(y k )*x k *2 -k
[0069] z k+1 =z k -sign(y k )*tan -1 (2 -k )
[0070] where tan -1 (1) and tan -1 (2 -k) is stored in the lookup table. In this embodiment, each CORDIC module operator is iterated 23 times to complete the result convergence. The number of iterations matches the pipeline level, and finally a high-throughput fixed-point implementation of 2 to 10 square root calculations can be completed. Software simulation confirms that the maximum number of positive iterations selected is 20, which is sufficient to meet the requirements of good accuracy. At the same time, the maximum index of negative iterations is -2. For the three coordinate systems, the negative expansion method is slightly different, specifically:
[0071] (a) For the circular coordinate system CORDIC algorithm, the sequence given by k = 0, 0, 0, 1, ..., 20 has been verified to be a better sequence than k = -2, -1, 0, 1, ..., 20;
[0072] (b) For the linear coordinate system CORDIC algorithm, the iteration index set is expanded to k = -2, -1, 0, 1, ..., 20, as shown in Figure 3 Show;
[0073] (c) For the hyperbolic coordinate system CORDIC algorithm, as with the linear coordinate system CORDIC algorithm, the iteration index set is expanded to k = -2, -1, 0, 1, ..., 20. The difference from the linear coordinate system CORDIC algorithm is that in the hyperbolic coordinate system CORDIC algorithm, the negative iteration has an iteration formula independent of the positive iteration operation.
[0074] This embodiment expands the complex input range through negative iteration, and performs simulation calculations of 10,000 data through MATLAB. In this embodiment, the number of positive iterations determines the accuracy, and the negative iteration is used to expand the convergence range, which has great flexibility.
[0075] Assuming that the maximum number of negative iterations is -2 and the maximum number of positive iterations is P, on this basis, the relationship between the average relative error and the integer N value and the positive iteration number P is explored in the experiment.
[0076] The integer N supports values from 2 to 10. When P varies from 15 to 20, we get Figure 4 The simulation relationship diagram of the average relative error, integer N value and positive iteration number P is shown. When P = 20, the average relative error can reach 1.38*10 -6 , can support 10 -8 to 10 4 The input range is
[0077] This embodiment uses Verilog HDL language for modeling and performs hardware simulation based on 10,000 test data. The average relative error is 2.9578*10 -6Taking N=10 as an example, an example circuit using TSMC 28nm CMOS technology is synthesized through hardware implementation. Table 2 is a comprehensive result and performance comparison table of this embodiment and the prior art.
[0078] Table 2
[0079] Technology Architecture frequency <![CDATA[Area (μm 2 )]]> Calculating Delay Accuracy Preliminary work 1 28nm Non-pipelined architecture 1.5GHz 6561 170.9ns <![CDATA[9.6117*10 -5 ]]> Preliminary work 2 28nm Pipeline architecture 2.218GHz 67964.17 4.50ns <![CDATA[2.9660*10 -6 ]]> The present invention works 28nm Pipeline architecture 2.218GHz 68070.87 0.451ns <![CDATA[2.9578*10 -6 ]]>
[0080] Preliminary work 1 is a patent cited in the background technology and is the result of the applicant's preliminary work. This method operates at a frequency of 1.5 GHz and does not adopt a pipeline architecture. It can only serially calculate the Nth root of each input complex number based on the value of N, and only calculates one of the N results each time. Taking N=10 as an example, calculating the 10th root of a complex number requires 255 clock cycles, or 170.9ns.
[0081] Preliminary work 2 is to add a pipeline architecture based on preliminary work 1, with an operating frequency of 2.218 GHz. Taking N=10 as an example, only 1 of the 10 roots is calculated each time, and it takes 10 clock cycles to calculate the 10th root of a complex number, which takes 4.5ns.
[0082] Based on the previous work 2, this embodiment designs a parallel result processing unit and uses the similarity of the N-root solution process of a complex number to reduce the computational complexity. Taking N=10 as an example, after earlier calculations, it only takes one clock cycle, or 0.45ns, to calculate the 10th root of a complex number.
[0083] According to the comprehensive results in Table 2, the circuit complexity of this embodiment is only 0.157% higher than that of the previous work 2, but the calculation speed is increased by 10 times. Compared with the previous work 1, the calculation efficiency of the system in this embodiment can be increased by 379 times at most, and the hardware implementation accuracy is also improved by 1 order of magnitude.
[0084] In summary, the present invention provides a low hardware complexity architecture that dynamically supports complex number 2 to 10th root operations. The architecture uses three CORDIC systems: circular, linear, and hyperbolic to build a hardware efficient algorithm. The convergence range is expanded, and software simulation can support 10 -8 to 10 4 The input range reaches 10 -6 The relative error of the algorithm is large, and it supports a wide range and high precision. It adopts a pipeline architecture to achieve high-speed full-pipeline calculation. It uses the similarity of the N-th root solution process of complex numbers to reduce the calculation complexity. It can dynamically support the calculation of the 2nd to 10th roots of complex numbers, and has the characteristics of high efficiency, high precision and low hardware complexity.
[0085] The invention and its implementation methods are described schematically above, and the description is non-restrictive. Although the invention has been represented and described with reference to specific preferred embodiments, it shall not be construed as a limitation on the invention itself. Various changes may be made to it in form and detail without departing from the spirit or basic characteristics of the invention. Without departing from the purpose of the invention, structural methods and embodiments similar to the technical solution are designed without creativity, and all of them shall belong to the protection scope of the present application. The multiple elements stated in the present application may also be implemented by one element through software or hardware. Words such as first and second are used to represent names, and do not represent any specific order.
Claims
1. A low-complexity hardware system based on CORDIC, characterized in that: It includes a circular vector mode CV-CORIDC module, a circular rotation mode CR-CORDIC module, a first linear vector mode LV1-CORDIC module, a second linear vector mode LV2-CORDIC module, a hyperbolic vector mode HV-CORDIC module, a hyperbolic rotation mode HR-CORDIC module and a result processing unit; The input data of the system is input to the input end of the CV-CORIDC module, the output end of the CV-CORIDC module is connected to the input end of the HV-CORDIC module and the input end of the LV2-CORDIC module, the output end of the HV-CORDIC module is connected to the input end of the LV1-CORDIC module, the output end of the LV1-CORDIC module is connected to the input end of the HR-CORDIC module, the output end of the HR-CORDIC module and the output end of the LV2-CORDIC module are both connected to the input end of the CR-CORDIC module, the output end of the CR-CORDIC module is connected to the input end of the result processing unit, and the result processing unit outputs the calculation result; The result processing unit includes a plurality of parallel result units. The result processing unit dynamically activates different result units according to different input integers N, and completes the calculation of N roots in parallel.
2. A low-complexity hardware system based on CORDIC according to claim 1, characterized in that: The system uses a pipeline architecture.
3. A method for calculating the 2nd to 10th root of a complex number, characterized in that: A low-complexity CORDIC-based hardware system as described in claim 1 or 2 is used; the expression of the complex number z is: z=m+j*n, and the Nth root of the complex number z is calculated by the exponential form of the complex number, where N is an integer greater than or equal to 2 and less than or equal to 10.
4. A complex number 2 to 10th root calculation method according to claim 3, characterized in that: The result processing unit performs shift and addition operations, and obtains the sum and difference formula of the two angles of trigonometric functions. The real and imaginary parts of .
5. A complex number 2 to 10th root calculation method according to claim 4, characterized in that: According to N different d values, There are N roots, d = 0, 1, ..., N-1; when the value of d is determined, and is a constant that is calculated in advance and stored in a lookup table.
6. A method for calculating the 2nd to 10th root of a complex number according to claim 4, characterized in that: The integer N value entered in each clock cycle changes dynamically between 2 and 10.
7. A complex number 2 to 10th root calculation method according to claim 4, characterized in that: Each CORDIC module is iterated X times through an X-stage pipeline architecture to complete result convergence, where the number of iterations includes a positive number of iterations for determining calculation accuracy and a negative number of iterations for expanding the calculation convergence range, and X is an integer greater than zero.
8. A complex number 2 to 10th root calculation method according to claim 5, characterized in that: The calculated constant values related to N are stored in a lookup table, and the corresponding calculated values are obtained according to different integers N between 2 and 10.
