WO2023173592A1 - 基于间断有限元法的集成电路互连线寄生电容提取方法 - Google Patents

基于间断有限元法的集成电路互连线寄生电容提取方法 Download PDF

Info

Publication number
WO2023173592A1
WO2023173592A1 PCT/CN2022/095540 CN2022095540W WO2023173592A1 WO 2023173592 A1 WO2023173592 A1 WO 2023173592A1 CN 2022095540 W CN2022095540 W CN 2022095540W WO 2023173592 A1 WO2023173592 A1 WO 2023173592A1
Authority
WO
WIPO (PCT)
Prior art keywords
boundary
unit
grid unit
grid
adjacent
Prior art date
Application number
PCT/CN2022/095540
Other languages
English (en)
French (fr)
Inventor
蔡志匡
杨航
赵郑豪
朱洪强
王恒鹭
郭静静
姚佳飞
郭宇锋
Original Assignee
南京邮电大学
南京邮电大学南通研究院有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 南京邮电大学, 南京邮电大学南通研究院有限公司 filed Critical 南京邮电大学
Publication of WO2023173592A1 publication Critical patent/WO2023173592A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/30Circuit design
    • G06F30/39Circuit design at the physical level
    • G06F30/398Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM]

Definitions

  • the invention relates to integrated circuit design, specifically discloses a method for extracting parasitic capacitance of integrated circuit interconnection lines based on the discontinuous finite element method, and belongs to the technical field of calculation, calculation or counting.
  • circuit optimization design has become an important stage in the integrated circuit design process.
  • the purpose of circuit optimization is to improve the electrical performance of the circuit.
  • the final actual electrical performance of the circuit not only depends on the device parameter values of the circuit, but also depends on the parasitic effects of the device itself, the parasitic effects between devices, the parasitic effects of the connection itself, and the connection.
  • the parasitic effects between lines, as well as the parasitic effects between connections and devices, and the parasitic effects between adjacent connections are particularly critical. From the perspective of circuit optimization theory, in order to obtain accurate circuit optimization results, it is necessary to accurately consider the parasitic effects between the connections of various devices on the designed circuit, especially the parasitic effects between capacitors, that is, parasitic capacitance.
  • the present invention aims to propose a method for extracting parasitic capacitance of integrated circuit interconnection lines based on the discontinuous finite element method.
  • the purpose of the present invention is to address the deficiencies of the above background technology and provide a method for extracting parasitic capacitance of integrated circuit interconnection lines based on the discontinuous finite element method.
  • the accuracy required by the industry standard can be achieved.
  • the running time it solves the technical problem that the existing interconnection line parasitic capacitance extraction method needs to sacrifice one indicator of extraction accuracy and running time to meet the performance improvement requirements of the other indicator.
  • a method for extracting parasitic capacitance of integrated circuit interconnection lines based on discontinuous finite element method including steps S1 to S6.
  • S1 For any given conductor distribution, first read the coordinates of the lower left point and the upper right point of the integrated circuit layout area. For each conductor, use several cuboids to approximately fit the shape of the conductor, and read the coordinates of each cuboid. The coordinates of the lower left point and the upper right point are recorded and numbered. All point coordinates are read and classified into row and column coordinates. The gaps between the row (or column) coordinates of each two adjacent conductors are divided into intervals: first, the rows close to the boundary of the two conductors are (or column) is expanded outward several times on average with a very small spacing distance of mingap.
  • the spacing distance is gradually expanded by a multiple of gaptime and then unequal division is performed until the last two expansions.
  • the distance between adjacent conductor boundary rows (or columns) is smaller than the enlarged interval again, forming a grid division that is uniform and dense when close to the conductor boundary, and uneven and sparse when far away from the conductor boundary.
  • S3 Traverse all global units from scratch. If the flag of the unit indicates that it is a boundary unit, then we need to pre-given the potential degree of freedom of the unit. Generally, for the boundary of the required solution area and the boundary of the non-main conductor, We assign all potential degrees of freedom to 0; if the boundary unit is the main conductor boundary, then we set the constant term of the potential degree of freedom of the boundary unit to 1, and set the constant terms of the potential degrees of freedom of the other boundary units to 0.
  • S4 Traverse all the grid units that need to be solved from scratch until the unit number that needs to be solved is traversed: for each grid unit, first analyze whether its neighboring units are boundary units, and get the difference based on whether the neighboring units are boundaries or not.
  • the electrostatic field strength equation of Six cells the electrostatic field strength equation of the LDG format is given based on the boundaries of these six grid cells.
  • each grid unit can be calculated from the electric potential degree of freedom, after obtaining the electric potential degree of freedom, the electric field strength degree of freedom can be calculated and the x-axis direction electric field strength sum of each grid unit can be obtained.
  • the four leftmost, rightmost, topmost and bottommost sides are regarded as the four side lengths of the conductor. These four sides extend one grid as a Gaussian surface and are integrated.
  • the discontinuous finite element method proposed by this invention inherits all the advantages of the finite element method, and proposes a non-uniform grid division method of close and far sparse, which improves the solution accuracy by densifying the grid near the boundary of the conductor unit.
  • a set of basis functions is taken for each discontinuous grid unit during the calculation process to avoid
  • the traditional finite element method realizes precision adaptation, the basis function continuously appears hanging points within a certain range.
  • the extraction accuracy required by industry standards is obtained. Compared with the traditional finite element method, it has the technical advantages of improving the extraction accuracy of parasitic capacitance of integrated circuit interconnection lines, shortening the running time, reducing the running memory, and easily realizing adaptive adjustment of accuracy.
  • the discontinuous finite element method proposed by the present invention has high accuracy, and only needs to modify one accuracy parameter in the program to easily obtain the desired accuracy, and the meshing parameters can be slightly modified to achieve the desired accuracy. It can well cope with various complex boundary conditions and meshing. Compared with the finite difference method, it has the technical advantages of flexible meshing, adaptive adjustment of extraction accuracy and short running time.
  • Figure 1 is an equivalent circuit diagram of the electromagnetic coupling effect of an integrated circuit.
  • Figure 2 is a flow chart of a method for extracting parasitic capacitance of interconnect lines provided by the present invention.
  • Figure 3(a) is a schematic diagram of the meshing between the boundaries of two adjacent conductors
  • Figure 3(b) is a schematic diagram of the meshing of the local solution area.
  • Figure 4(a) is a schematic diagram of the boundary of grid unit k
  • Figure 4(b) is a schematic diagram of the potential and electric field intensity flux on the boundary of grid unit k.
  • the integrated circuit as shown in Figure 1 includes: a main conductor 1, a second conductor 2 adjacent to the main conductor, and a third conductor 3 adjacent to the main conductor.
  • the main conductor 1, the second conductor 2, and the third conductor 3 The potentials on The capacitances of the main conductor 1, the second conductor 2, and the third conductor 3 are C 20 , C 10 , and C 30 respectively.
  • the parasitic capacitances of the interconnection lines between the main conductor 1 and the second conductor 2 are C 12 , the main conductor 1, and the third conductor 3 .
  • the parasitic capacitance of the interconnection line between the third conductor 3 is C 13
  • the parasitic capacitance of the interconnection line between the second conductor 2 and the third conductor 3 is C 23 .
  • the present invention provides a method for extracting parasitic capacitance of integrated circuit interconnection lines based on discontinuous finite element generation, including the following six steps.
  • Step 1 After reading the data of each conductor, divide the integrated circuit layout into non-uniform meshes.
  • the spacing distance is gradually expanded by a multiple of gaptime and then unequal division is performed until the last two expansions.
  • the distance between rows (or columns) of adjacent conductor boundaries is smaller than the enlarged interval again, forming a grid division that is uniform and dense when close to the conductor boundary, and uneven and sparse when far away from the conductor boundary.
  • the boundary between two adjacent conductors The grid is shown in Figure 3(a).
  • Step 2 Set the boundaries and numbers of grid cells
  • the required solution area is obtained by combining fine rectangular grid units. It is judged whether the grid unit is a boundary unit and the lateral direction of each grid unit is obtained. dx length and longitudinal dy length, each grid unit is marked with a global number.
  • the present invention defines grid units falling outside the boundary of the required solution area and grid units within the area where the conductor is located as boundary units, and the local required solution area
  • the grid division is shown in Figure 3(b); each grid unit is traversed, and those grid units that are not boundaries are assigned a second number, recorded as "solving unit number", and a global number is established to the solving unit Number mapping relationship.
  • Step 3 Initialize the degrees of freedom of all grid cells, obtain the electrostatic field strength formula in the LDG format based on whether the adjacent grid cells of the required grid unit are boundary cells, and solve for the potential function freedom of all the grid cells that need to be resolved. degrees of freedom and electric field intensity function
  • Electrostatic field strength formula For each grid unit that needs to be solved, we multiply the electric potential function u(x, y), electric field strength function p(x, y), q(x, y) on each grid unit that needs to be solved by each The basis function of each grid unit that needs to be solved (equivalent to finding the projection of the electric potential function and the electric field strength function on the basis function of each grid unit that needs to be solved), and the integration of the grid unit that needs to be solved here is expressed by the following formula Electrostatic field strength formula:
  • K xw and K xe represent the left and right boundaries of grid unit k
  • K ys and K yn represent the lower and upper boundaries of grid unit k.
  • the four boundaries of grid unit k are shown in Figure 4(a)
  • (Kx0, Ky0) represents the center coordinate of grid unit k; represents numerical flux, Represents the numerical flux of electric field strength q(x, -1/2) at the lower boundary of grid unit k, q(x, -1/2) represents the electric field strength at the lower boundary of grid unit k, Represents the numerical flux of electric field strength q(x, 1/2) at the upper boundary of grid unit k, q(x, 1/2) represents the electric field strength at the upper boundary of grid unit k, Represents the numerical flux of electric field strength p(-1/2, y) at the left boundary of grid unit k, p(-1/2, y) represents the electric field strength at the left boundary of grid unit k, Represents the numerical flux of electric field strength p(
  • p, q, and u on the upper and right sides of the grid unit boundary are given symbols p + , q + , u + , and p on the lower and left side of the grid unit boundary, q and u are given symbols p - , q - , u - , and the actual potential of the boundary is u exact .
  • module one when K n and K nn (that is, the upper boundary unit and the upper boundary unit of grid unit k) are not boundary units, module one looks like:
  • the module has the following shape:
  • module two when K n is not a boundary unit, module two looks like:
  • module 2 looks like:
  • module three when K n is not a boundary unit, module three looks like:
  • module three looks like:
  • module four when neither K e nor K ee is a boundary unit, module four looks like:
  • module four looks like:
  • module five when K e is not a boundary unit, module five looks like:
  • module five looks like:
  • module six when K e is not a boundary element, module six looks like:
  • module six looks like:
  • ⁇ x k represents the length of grid unit k in the x direction
  • ⁇ y k represents the length of grid unit k in the y direction
  • ⁇ y kn represents the length of the adjacent grid unit above grid unit k in the y direction
  • k potential function degree of freedom vector of the grid unit represents the degree of freedom vector of the potential function of the grid unit adjacent to the upper boundary of grid unit k
  • ⁇ y kn represents the length of freedom vector of the potential function of the grid unit adjacent to the lower boundary of grid unit k
  • the degree of freedom vector of the potential function of the grid unit adjacent to the upper boundary of the upper boundary grid unit of grid unit k, ⁇ x ke represents the length of the adjacent grid unit to the right of grid unit k in the x direction
  • the first step is to assign a value to the potential degree of freedom of the boundary unit. If it is not the main conductor boundary unit, then the potential degree of freedom on the boundary unit is 0. If the boundary unit is the main conductor Boundary unit, then assign the constant value term of the potential degree of freedom of the boundary unit to 1, and assign the value of the other boundary unit potential degree of freedom terms to 0; in the second step, set the right-hand term of the linear equation system that needs to be solved, and set them all to 0; The third step is to traverse the solution unit number from scratch, perform boundary judgment for each grid unit k that needs to be solved, and determine whether its Kn, Knn, Ks, Ke, Kee, Kw units are boundary units in turn, and based on the obtained boundaries Conditions (i.e., the judgment results of whether Kn, Knn, Ks, Ke, Kee, Kw units are boundary units) perform different processing on modules 1, 2, 3, 4, 5 and 6 in the electrostatic field strength formula respectively, and
  • Step 4 Solve the charge of each conductor and the coupling capacitance of each conductor based on the electric field strength of each grid unit
  • the Gaussian integral of each conductor can be calculated to obtain the charge size of each conductor, and then the main conductor capacitance and the coupling capacitance of each conductor can be obtained.
  • the solution method is as follows: since each rectangular grid unit has been numbered before, all rectangular grid units are traversed, all rectangular grid units with the same number are regarded as the same conductor, and all rectangular grid units are taken The leftmost, rightmost, uppermost, and lowermost boundaries of the unit are used as the four boundaries of the conductor.
  • the grid unit set composed of the grid units enclosed by the four boundaries and the outer grid unit of the four boundaries is used to solve the conductor.
  • each grid unit in the grid unit set is separately integrated with the electric field intensity function in the x-axis direction and the electric field intensity function in the y-axis direction.
  • Each grid unit k on the Gaussian surface is integrated with the x-axis
  • the integral of the electric field intensity function in the direction has the following integral form:
  • Kx0 represents the x coordinate of the midpoint of grid unit k
  • Ky0 represents the y coordinate of the midpoint of grid unit k
  • ⁇ x k represents the length of grid unit k in the x direction
  • ⁇ y k represents the length of grid unit k in the y direction.
  • the length on, K xw , K xe represents the left and right boundaries of grid unit k
  • q l represents the degree of freedom of the electric field strength function in the y direction on grid unit k.
  • the electric field intensity function in the y-axis direction is integrated for each grid unit k on the Gaussian surface.
  • the integral form is as follows:
  • K ys and K yn represent the lower and upper boundaries of grid unit k
  • p l represents the degree of freedom of the electric field intensity function in the y direction on grid unit k.
  • the present invention tested the accurate solution of the proposed algorithm and the calculation example of the error order.
  • the accuracy was tested, and it was finally found that the error of the test results met the error requirements of the theoretical calculation.
  • the second-order error accuracy test results with conductors are shown in Table 1, and the second-order error accuracy test results with no conductors are shown in Table 2. It can be seen that the discontinuous finite element algorithm of the present invention is reliable and the code is correct. of.
  • the verification results of the extraction accuracy and running time of the parasitic capacitance between different numbers of conductors by the discontinuous finite element algorithm of the present invention are as follows: For an example where the number of conductors is 3, the error between the numerical result of the parasitic capacitance extracted by the present invention and the standard reference value The size is 0.65%, and the running time is 0.16 seconds; for the example where the number of conductors is 8, the error between the numerical result of the parasitic capacitance extracted by the present invention and the standard reference value is 0.52%, and the running time is 0.52 seconds; for the individual conductors; For an example where the number of conductors is 71, the error between the numerical result of the parasitic capacitance extracted by the present invention and the standard reference value is 1.63%, and the running time is 2.3 seconds; for an example where the number of conductors is 97, the error of the parasitic capacitance extracted by the present invention is The error between the numerical results and the standard reference value is 2.40%, and the running time is 3.1 seconds.
  • the error of the parasitic capacitance extracted by the present invention is smaller.
  • the error of the parasitic capacitance extracted by the present invention is still not large, and is controlled within 2.5%. And it runs very fast.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Design And Manufacture Of Integrated Circuits (AREA)

Abstract

本发明公开基于间断有限元法的集成电路互连线寄生电容提取方法,属于计算、推算或计数的技术领域。该方法根据导体分布情况划分非均匀的矩形网格;判断矩形网格是否为边界单元格,依次标录全局编号和需求解编号;初始化所有矩形网格的自由度值;遍历所有矩形网格,根据邻近单元网格的边界情况,得到基于间断有限元法的线性方程组,计算所有矩形网格电势函数自由度;根据每个矩形网格的电势函数自由度求出每个单元的电场强函数自由度;划分出每个导体的高斯面,在高斯面上积分求得的电场强函数得到电荷,最终求出主导体电容以及耦合电容大小。本发明提高集成电路互连线寄生电容提取精度,并且降低运行时间及运行内存。

Description

基于间断有限元法的集成电路互连线寄生电容提取方法 技术领域
本发明涉及集成电路设计,具体公开基于间断有限元法的集成电路互连线寄生电容提取方法,属于计算、推算或计数的技术领域。
背景技术
随着科技的发展,电路优化设计成为集成电路设计流程中的一个重要阶段。电路优化的目的就是提高电路的电学性能,而电路的最终实际电学性能不仅取决电路的器件参数值,还取决于器件本身的寄生效应、器件之间的寄生效应、连线本身的寄生效应、连线之间的寄生效应、以及连线和器件之间的寄生效应,而相邻连线之间的寄生效应尤为关键。从电路优化理论上来讲,为了得到准确的电路优化结果,需要精确考虑所设计的电路上的各个器件连线之间的寄生效应,尤其是电容之间所产生的寄生效应,即寄生电容。
随着半导体工艺制程的发展,如何准确快速地进行寄生电容的提取对于保证芯片设计质量,满足严苛的PPA指标要求,缩短设计周期变得至关重要。互连线电容计算是集成电路EDA(Electronic Design Automation,电子设计自动化)领域经典的求解课题和主流诉求,算法创新与优化永无止尽,是一个持续的研究热点。
寄生参数提取早期使用的解析方法只适合于一些非常简单的互连结构,随着互连结构的日益复杂以及求解精度要求的逐步提高,直接场求解成为一种有效的互连线寄生电容提取方法,但对于庞大的互连结构,全部都用场来直接求解所需时间多且所需运行内存大,不能满足大规模集成电路EDA的需求。有限差分法和有限元法等区域离散化方法可以解决直接场求解法处理庞大互联结构时间开销大的问题,但有限差分法对网格的划分要求较为严格,且精度有待提高,而有限元法在实现精度自适应调整时对悬点的处理比较棘手。为此,亟待需要一种精度高、运行时间短以及运行内存小的集成电路互连线之间寄生电容的提取方法。本发明旨在提出一种基于间断有限元法的集成电路互连线寄生电容提取方法。
发明内容
本发明的发明目的是针对上述背景技术的不足,提供基于间断有限元法的 集成电路互连线寄生电容提取方法,通过对有限元法的网格进行非均匀划分,在满足行业标准要求的精度的同时缩短运行时间,解决现有互连线寄生电容提取方法需牺牲提取精度与运行时间中的一个指标以满足另一指标性能提升要求的技术问题。
本发明为实现上述发明目的采用如下技术方案:
一种基于间断有限元法的集成电路互连线寄生电容的提取方法,包括步骤S1至步骤S6。
S1:对于任意给定的导体分布情况,首先读取集成电路版图区域的左下点坐标和右上点坐标,对于每个导体,用若干个长方体去近似拟合导体的形状,读取每个长方体的左下点和右上点坐标并记录编号,读取所有的点坐标并分类为行列坐标,对每两个相邻导体的行(或列)坐标中间的空隙进行间隔划分:首先在靠近两导体边界行(或列)处以一个极小的值mingap的间隔距离平均向外扩展若干次,在进行完若干次等距划分后,以gaptime的倍数逐步扩大间隔距离后进行不等距划分,直至扩展后两个相邻导体边界行(或列)的距离小于再次扩大后的间隔,形成靠近导体边界时均匀且密,远离导体边界时不均匀且疏的网格划分。
S2:在得到被划分成由细小矩形单元组合而成的需求解区域后,对每个网格单元的信息进行计算,判断每个网格单元是否是边界单元(用一个flag标识标记),求出每个网格单元的横向dx长度和纵向dy长度,并对每个网格单元标注一个全局编号;遍历所有网格单元,对那些不是边界的网格单元赋予第二个编号,记作“求解单元编号”,并建立全局编号到求解单元编号的映射关系。
S3:从零开始遍历所有的全局单元,如果该单元的flag标识表明其为边界单元,那么我们需要预先给定该单元的电势自由度,一般地,对于需求解区域边界和非主导体边界,我们赋值所有电势自由度为0;如若该边界单元为主导体边界,那么我们把该边界单元电势自由度的常数项设为1,其余边界单元电势自由度的常数项设为0。
S4:从零开始遍历所有需要求解的网格单元,直至遍历完需求解单元编号:对于每一个网格单元首先分析它的邻近单元是否为边界单元,根据其邻近单元的边界与否来得到不同的静电场场强等式,具体算法用到间断有限元法的LDG格式,以某一个具体网格单元k为例,需要分析它邻近的网格单元kn,Knn,Ks,Ke,Kee,Kw六个单元,以这6个网格单元的边界与否来给出LDG格式 的静电场强等式。
S5:将分析所有需求解网格单元边界后推出的静电场场强等式组合成一个
Figure PCTCN2022095540-appb-000001
Figure PCTCN2022095540-appb-000002
形式的线性方程组,并求解该线性方程组得到每个需求解单元的电势自由度。
S6.由于每个网格单元的电场强自由度可以由电势自由度求出,因此,在得到电势自由度后可以计算得到电场强自由度并得到每一个网格单元的x轴方向电场强和y轴方向的电场强;在S1中,我们给每个长方体编了号,现在我们从零开始遍历所有长方体,把所有编号相同的长方体视作同一个导体,把每个导体中的所有长方体中的最左最右最上和最下的四条边视作是该导体的四条边长,对这四条边外延一格作为高斯面进行积分,具体地,对上边和下边乘上y方向的电场强后求积,对左边和右边乘上x方向的电场强后进行求积,将所有数值加起来得到具体一个导体的电荷量,并通过电荷量和电容大小的关系解算出主导体的电容大小和其余导体的耦合电容大小。
本发明采用上述技术方案,具有以下有益效果:
(1)本发明提出的间断有限元法,继承了有限元法的所有优势,提出了一种近密远疏的非均匀网格划分方式,通过加密导体单元边界附近的网格提高求解精度,以解决求解问题中边界区域附近的数值解产生奇异性的问题,并且通过修正网格划分的参数实现局部网格自适应,计算过程中给相互间断的网格单元分别取一组基函数,避免传统有限元法实现精度自适应时因基函数在某一范围内连续出现悬点的问题,采用本发明提出的间断有限元法提取不同规模的集成电路的寄生电容时,以较少的运行时间得到行业标准要求的提取精度,相较于传统有限元法具有提高集成电路互连线寄生电容提取精度、缩短运行时间、降低运行内存以及易于实现精度自适应调整的技术优势。
(2)本发明提出的间断有限元法的精度高,且只需要修改程序中的一个精度参数就可以很方便地得到所想要的精度大小,且可以对网格划分参数稍作修改就很好地应对各种复杂边界情况和网格划分,相较于有限差分法具有网格划分灵活、提取精度自适应调整且运行时间短的技术优势。
附图说明
图1是集成电路电磁耦合效应的等效电路图。
图2是本发明提供的一种提取互连线寄生电容方法的流程图。
图3(a)是相邻两导体边界之间的网格划分示意图,图3(b)为局部需求解区域的网格划分示意图。
图4(a)是网格单元k的边界的示意图,图4(b)是网格单元k边界上电势及电场强流通量的示意图。
具体实施方式
为了使本技术领域的人员更好地理解本发明实施例的方案,下面结合附图和实施方式对本发明实施例作进一步的详细说明。
如图1所示的集成电路,包括:主导体1、与主导体相邻的第二导体2、与主导体相邻的第三导体3,主导体1、第二导体2、第三导体3上的电势分别为
Figure PCTCN2022095540-appb-000003
主导体1、第二导体2、第三导体3的电容分别为C 20、C 10、C 30,主导体1、第二导体2之间互连线的寄生电容为C 12、主导体1、第三导体3之间互连线的寄生电容为C 13,第二导体2、第三导体3之间互连线的寄生电容为C 23
如图2所示,本发明提供的一种基于间断有限元发的集成电路互连线寄生电容的提取方法,包括以下六大步骤。
步骤一:读取各导体的数据后对集成电路版图进行非均匀网格划分
对于任意给定的导体分布情况,首先读取各导体所占版图区域的左下点坐标和右上点坐标,对于每个导体,用若干个长方体去近似拟合导体的形状,读取每个长方体的左下点和右上点坐标并记录编号,将读取的所有的点坐标分类为行列坐标,对每两个相邻导体的行(或列)坐标中间的空隙进行间隔划分:首先在靠近两导体边界行(或列)处以一个极小值mingap的间隔距离平均向外扩展若干次,在进行完若干次等距划分后,以gaptime的倍数逐步扩大间隔距离后进行不等距划分,直至扩展后两个相邻导体边界行(或列)的距离小于再次扩大后的间隔,形成靠近导体边界时均匀且密,远离导体边界时不均匀且疏的网格划分,相邻两个导体之间的边界网格如图3(a)所示。
步骤二:设置网格单元的边界性与编号
对集成电路版图中各导体之间的间隔进行上述非均匀划分后得到由细小矩形网格单元组合而成的需求解区域,判断网格单元是否是边界单元并求出每个网格单元的横向dx长度和纵向dy长度,对每个网格单元标注一个全局编号,本发明定义落在需求解区域边界外的网格单元以及导体所在区域内的网格单元为边界单元,局部需求解区域的网格划分如图3(b)所示;对每个网格单元进行遍 历,对那些不是边界的网格单元赋予第二个编号,记作“求解单元编号”,并建立全局编号到求解单元编号的映射关系。
步骤三:初始化所有网格单元的自由度,根据需求解网格单元的相邻网格单元是否为边界单元的情形得到LDG格式的静电场强公式,求解所有需求解网格单元的电势函数自由度及电场强函数自由度
将需求解网格单元里每个点的电势大小记作u(x,y),每个点的x方向的电场强大小记作p(x,y),每个点的y方向的电场强大小记作q(x,y),即:
u x(x,y)=p(x,y),u y(x,y)=q(x,y)
于是原来的静电场场强公式Δu(x,y)=0就可以写作p x+q y=0的形式;再在每一个网格单元k上取一组基函数
Figure PCTCN2022095540-appb-000004
如此,我们便可以用
Figure PCTCN2022095540-appb-000005
的形式来近似地表示每个网格单元上的电势函数与电场强函数,其中,u l,p l,q l称为每个网格单元上电势函数和电场强函数的自由度。对于每个需求解的网格单元,我们令每个需求解网格单元上的电势函数u(x,y),电场强函数p(x,y),q(x,y)分别乘上每个需求解网格单元的基函数(相当于求电势函数以及电场强函数在每个需求解网格单元的基函数上的投影),并在此需求解网格单元进行积分,采用如下公式表示静电场场强公式:
Figure PCTCN2022095540-appb-000006
Figure PCTCN2022095540-appb-000007
Figure PCTCN2022095540-appb-000008
得到如式(1),(2),(3)所示的三个LDG(Local Discontinuous Galerkin,局部间断有限元法)格式的公式:
Figure PCTCN2022095540-appb-000009
Figure PCTCN2022095540-appb-000010
Figure PCTCN2022095540-appb-000011
其中,K xw,K xe表示网格单元k的左边界和右边界,K ys,K yn表示网格单元k的下边界和上边界,网格单元k的四个边界如图4(a)所示,(Kx0,Ky0)表示网格单元k的中心坐标;
Figure PCTCN2022095540-appb-000012
表示数值通量,
Figure PCTCN2022095540-appb-000013
表示电场强q(x,-1/2)在网格单元k下边界的数值流通量大小,q(x,-1/2)表示网格单元k下边界处的电场强,
Figure PCTCN2022095540-appb-000014
表示电场强q(x,1/2)在网格单元k上边界处的数值流通量大小,q(x,1/2)表示网格单元k上边界处的电场强,
Figure PCTCN2022095540-appb-000015
表示电场强p(-1/2,y)在网格单元k左边界处的数值流通量大小,p(-1/2,y)表示网格单元k左边界处的电场强,
Figure PCTCN2022095540-appb-000016
表示电场强p(1/2,y)在网格单元k右边界处的数值流通量大小,p(1/2,y)表示网格单元k右边界处的电场强,
Figure PCTCN2022095540-appb-000017
表示电势u(1/2,y)在网格单元k右边界处的数值流通量大小,u(1/2,y)表示网格单元k右边界处的电势,
Figure PCTCN2022095540-appb-000018
表示电势u(-1/2,y)在网格单元k左边界处的数值流通量大小,u(-1/2,y)表示网格单元k左边界处的电势,
Figure PCTCN2022095540-appb-000019
表示电势u(x,1/2)在网格单元k上边界处的数值流通量大小,u(x,1/2)表示网格单元k上边界处的电势,
Figure PCTCN2022095540-appb-000020
表示电势u(x,-1/2)在网格单元k下边界处的数值流通量大小,u(x,-1/2)表示网格单元k下边界处的电势;而
Figure PCTCN2022095540-appb-000021
表示基函数
Figure PCTCN2022095540-appb-000022
在网格单元k上边界的下方的流通量(即取值),
Figure PCTCN2022095540-appb-000023
表示网格单元k第l个基函数在上边界处的取值,
Figure PCTCN2022095540-appb-000024
表示基函数
Figure PCTCN2022095540-appb-000025
在网格单元k下边界的上方的流通量,
Figure PCTCN2022095540-appb-000026
Figure PCTCN2022095540-appb-000027
表示网格单元k第l个基函数在下边界处的取值,
Figure PCTCN2022095540-appb-000028
表示基函数
Figure PCTCN2022095540-appb-000029
在网格单元k右边界的左方的流通量,
Figure PCTCN2022095540-appb-000030
表示网格单元k第l个基函数在右边界处的取值,
Figure PCTCN2022095540-appb-000031
表示基函数
Figure PCTCN2022095540-appb-000032
在网格单元k左边界的右方的流通量,
Figure PCTCN2022095540-appb-000033
表示网格单元k第l个基函数在左边界处的取值。
如图4(b)所示,在网格单元边界上侧与右侧的p,q,u给与符号p +,q +,u +,在网格单元边界下侧与左侧的p,q,u给与符号p -,q -,u -,记边界的实际电势为u exact
当网格单元k的邻近网格单元不是边界单元时,
Figure PCTCN2022095540-appb-000034
当网格单元k的邻近网格单元为边界单元时:
Figure PCTCN2022095540-appb-000035
对于网格单元k左边界的p,有:
Figure PCTCN2022095540-appb-000036
对于网格单元k右边界上的p,有:
Figure PCTCN2022095540-appb-000037
对于网格单元k下边界的q,有:
Figure PCTCN2022095540-appb-000038
对于网格单元k上边界上的q,有:
Figure PCTCN2022095540-appb-000039
每个网格单元k以及其邻近网格单元K n,K s,K e,K w上的电势自由度合共产生了一组向量
Figure PCTCN2022095540-appb-000040
由式子(1),(2),(3),对于每个需求解的网格单元k,我们都可以列出形如
Figure PCTCN2022095540-appb-000041
的式子;遍历所有需求解的网格单元,将所有需求解网格单元的等式组合成一个大的形如Au=b线性方程组,便可以求出每个需求解网格单元上的电势自由度,从而求出各网格单元的电势大小并由此求出电场强大小。
但是由于每个网格单元需要考虑其邻近网格单元的边界性,因此我们将公式(1)分为6个模块,每个模块分别考虑其在边界处的处理条件。这6个模块分别为:
模块一:
Figure PCTCN2022095540-appb-000042
模块二:
Figure PCTCN2022095540-appb-000043
模块三:
Figure PCTCN2022095540-appb-000044
模块四:
Figure PCTCN2022095540-appb-000045
模块五:
Figure PCTCN2022095540-appb-000046
模块六:
Figure PCTCN2022095540-appb-000047
由于我们发现每个模块最后写成的矩阵向量相乘形式的系数矩阵可以由以下11个矩阵的加减乘运算得出,所以,首先我们引入参考网格单元与参考网格单元上的基函数的概念,所谓参考网格单元的基函数存在的意义就是,每个网格单元的基函数都可以通过某种变换把自身的形式变成参考网格单元上的基函数的形式,于是对于之后每个网格单元的基函数处理运算,我们都只需要在参考网格单元做处理和运算即可。
首先,我们定义参考网格单元上的基函数的符号为:
Figure PCTCN2022095540-appb-000048
参考网格单元的区域为[-1/2,1/2]×[-1/2,1/2]。
接下来,我们引入以下符号
Figure PCTCN2022095540-appb-000049
X nn、X ns、X sn、X ss、Y ee、Y ew、Y we、Y ww表示可以运算得出各模块矩阵相乘形式系数矩阵的11个矩阵:
Figure PCTCN2022095540-appb-000050
Figure PCTCN2022095540-appb-000051
Figure PCTCN2022095540-appb-000052
Figure PCTCN2022095540-appb-000053
Figure PCTCN2022095540-appb-000054
Figure PCTCN2022095540-appb-000055
Figure PCTCN2022095540-appb-000056
Figure PCTCN2022095540-appb-000057
Figure PCTCN2022095540-appb-000058
Figure PCTCN2022095540-appb-000059
Figure PCTCN2022095540-appb-000060
其中,
Figure PCTCN2022095540-appb-000061
表示参考网格单元K base第i个基函数
Figure PCTCN2022095540-appb-000062
在第j个基函数
Figure PCTCN2022095540-appb-000063
上的投影,
Figure PCTCN2022095540-appb-000064
表示参考网格单元K base第i个基函数x方向的导数在第j个基函数
Figure PCTCN2022095540-appb-000065
上的投影,
Figure PCTCN2022095540-appb-000066
表示参考网格单元K base第i个基函数y方向的导数在第j个基函数
Figure PCTCN2022095540-appb-000067
上的投影,
Figure PCTCN2022095540-appb-000068
表示参考网格单元K base第j个基函数在下边界取值
Figure PCTCN2022095540-appb-000069
与第i个基函数在下边界取值
Figure PCTCN2022095540-appb-000070
的乘积,
Figure PCTCN2022095540-appb-000071
表示参考网格单元K base第j个基函数在上边界取值
Figure PCTCN2022095540-appb-000072
与第i个基函数在下边界取值
Figure PCTCN2022095540-appb-000073
的乘积,
Figure PCTCN2022095540-appb-000074
表示参考网格单元K base第j个基函数在下边界取值
Figure PCTCN2022095540-appb-000075
与第i个基函数在上边界取值
Figure PCTCN2022095540-appb-000076
的乘积,
Figure PCTCN2022095540-appb-000077
表示参考网格单元K base第j个基函数在上边界取值
Figure PCTCN2022095540-appb-000078
与第i个基函数在上边界取值
Figure PCTCN2022095540-appb-000079
Figure PCTCN2022095540-appb-000080
的乘积,
Figure PCTCN2022095540-appb-000081
表示参考网格单元K base第j个基函数在左边界取值
Figure PCTCN2022095540-appb-000082
与第i个基函数在左边界取值
Figure PCTCN2022095540-appb-000083
的乘积,
Figure PCTCN2022095540-appb-000084
表示参考网格单元K base第j个基函数在右边界取值
Figure PCTCN2022095540-appb-000085
与第i个基函数在左边界取值
Figure PCTCN2022095540-appb-000086
的乘积,
Figure PCTCN2022095540-appb-000087
表示参考网格单元K base第j个基函数在左边界的取值
Figure PCTCN2022095540-appb-000088
与第i个基函数在右边界取值
Figure PCTCN2022095540-appb-000089
的乘积,
Figure PCTCN2022095540-appb-000090
表示参考网格单元K base第j个基函数在右边界取值
Figure PCTCN2022095540-appb-000091
与第i个基函数在右边界取值
Figure PCTCN2022095540-appb-000092
的乘积。
现在开始依次遍历需求解的网格单元,对于模块一,当K n,K nn(即网格单元k的上边界单元与上上边界单元)均不是边界单元时,模块一形如:
Figure PCTCN2022095540-appb-000093
当K n为非边界单元,但K nn为边界单元时,模块一形如:
Figure PCTCN2022095540-appb-000094
当K n为边界单元时,K nn的边界性不再产生影响(之后只在提到K nn与K ee时,才考虑K nn的边界性),于是模块一形如:
Figure PCTCN2022095540-appb-000095
对于模块二,当K n不是边界单元时,模块二形如:
Figure PCTCN2022095540-appb-000096
当K n是边界单元时,模块二形如:
Figure PCTCN2022095540-appb-000097
对于模块三,当K n不是边界单元时,模块三形如:
Figure PCTCN2022095540-appb-000098
当K n是边界单元时,模块三形如:
Figure PCTCN2022095540-appb-000099
对于模块四,当K e,K ee均不是边界单元时,模块四形如:
Figure PCTCN2022095540-appb-000100
当K e不是边界单元,但K ee是边界单元时,模块四形如:
Figure PCTCN2022095540-appb-000101
当K e为边界单元时,模块四形如:
Figure PCTCN2022095540-appb-000102
对于模块五,当K e不是边界单元,模块五形如:
Figure PCTCN2022095540-appb-000103
当K e是边界单元,模块五形如:
Figure PCTCN2022095540-appb-000104
对于模块六,当K e不是边界单元,模块六形如:
Figure PCTCN2022095540-appb-000105
当K e为边界单元时,模块六形如:
Figure PCTCN2022095540-appb-000106
Δx k表示网格单元k在x方向的长度,Δy k表示网格单元k在y方向上的长度,Δy kn表示网格单元k上方的邻近网格单元在y方向上的长度,
Figure PCTCN2022095540-appb-000107
表示网格单元k电势函数自由度向量,
Figure PCTCN2022095540-appb-000108
表示与网格单元k上边界相邻的网格单元的电势函数 自由度向量,
Figure PCTCN2022095540-appb-000109
表示与网格单元k下边界相邻的网格单元的电势函数自由度向量,
Figure PCTCN2022095540-appb-000110
与网格单元k上边界网格单元的上边界相邻的网格单元的电势函数自由度向量,Δx ke表示网格单元k右方的邻近网格单元在x方向上的长度,
Figure PCTCN2022095540-appb-000111
表示与网格单元k的右边界相邻的网格单元的电势函数自由度向量,
Figure PCTCN2022095540-appb-000112
表示与网格单元k右边界网格单元的右边界相邻的网格单元的电势函数自由度向量,
Figure PCTCN2022095540-appb-000113
表示与网格单元k左边界相邻的网格单元的电势函数自由度向量,
Figure PCTCN2022095540-appb-000114
表示与网格单元k右边界相邻的网格单元的电势函数自由度向量。
求解电势自由度的具体操作为:第一步,对边界单元的电势自由度进行赋值,如若不是主导体边界单元,那么该边界单元上的电势自由度均为0,如若该边界单元为主导体边界单元,那么对该边界单元电势自由度的常数值项赋值为1,其余边界单元电势自由度项赋值为0;第二步,设置需求解的线性方程组的右端项,全置为0;第三步,从零开始遍历求解单元编号,对于每个需求解网格单元k进行边界判断,依次判断其Kn,Knn,Ks,Ke,Kee,Kw单元是否为边界单元,并根据得到的边界条件(即Kn,Knn,Ks,Ke,Kee,Kw单元是否为边界单元的判断结果)分别对静电场场强公式中的模块一二三四五六做不同的处理,并把处理后得到的等式代入式(1),(2),(3)所示线性方程组中,最后得到一个
Figure PCTCN2022095540-appb-000115
形式的线性方程组,即为本发明基于间断有限元法,其中,A为一个阶数巨大的稀疏矩阵,
Figure PCTCN2022095540-appb-000116
为所有需求解网格单元电势自由度组合而成的电势函数自由度向量。
在求解出
Figure PCTCN2022095540-appb-000117
之后,可以根据公式(2),(3)依次求出
Figure PCTCN2022095540-appb-000118
Figure PCTCN2022095540-appb-000119
得到每个网格单元的电场强函数的自由度,但是由于有边界单元的存在,因此仍然需要考虑到边界条件,求出
Figure PCTCN2022095540-appb-000120
Figure PCTCN2022095540-appb-000121
之后就得到了每个网格单元关于x方向和y方向的电场强度。
步骤四:根据各网格单元的电场强度求解各导体的电荷以及各导体的耦合电容
得到每个网格单元的电场强度之后,就可以对每个导体求高斯积分来得到每个导体的电荷大小,进而求得主导体电容和各个导体的耦合电容。求解方法具体如下:由于之前已给每个矩形网格单元进行编号,因此对所有矩形网格单元进行遍历,将所有编号相同的矩形网格单元视为是同一个导体,并取所有矩形网格单元的最左、最右、最上、最下边界作为该导体的四条边界,取四条边界围合的网格单元以及这四条边界的外一格网格单元组成的网格单元集合作为求解该导体电容的高斯面,对网格单元集合中的各网格单元分别进行x轴方向电场强函数的 积分、y轴方向电场强函数的积分,对高斯面上的每个网格单元k进行x轴方向上电场强函数的积分,积分形式如下:
Figure PCTCN2022095540-appb-000122
其中,Kx0表示网格单元k中点的x坐标,Ky0表示网格单元k中点的y坐标,Δx k表示网格单元k在x方向上的长度,Δy k表示网格单元k在y方向上的长度,K xw,K xe表示网格单元k的左边界和右边界,q l表示网格单元k上的y方向的电场强函数的自由度。
对高斯面上的每个网格单元k进行y轴方向上电场强函数的积分,积分形式如下:
Figure PCTCN2022095540-appb-000123
其中,K ys,K yn表示网格单元k的下边界和上边界,p l表示网格单元k上的y方向的电场强函数的自由度。
累加高斯面上每个网格单元k的x轴方向上电场强函数积分值以及y轴方向上电场强函数的积分值,获得导体电荷量,依据导体电荷量与电容大小的关系求得各导体的电容大小以及导体之间互连线的寄生电容。
本发明测试了所提算法的准确解及误差阶的算例,选取了u=sin(x+y)这个例子,分别在有导体的情况和无导体的情况下对一阶误差和二阶误差的精度进行了测试,最终发现测试得到的结果的误差是满足理论计算的误差要求的。有导体情况下的二阶误差精度测试结果如表1所示,无导体情况下的二阶误差精度测试结果如表2所示,可见,本发明的间断有限元算法是可靠的,代码是正确的。
Figure PCTCN2022095540-appb-000124
表1
Figure PCTCN2022095540-appb-000125
Figure PCTCN2022095540-appb-000126
表2
本发明间断有限元算法对不同个数导体间的寄生电容的提取精度及运行时间的验证结果如下:对于导体个数为3个的例子,本发明提取的寄生电容数值结果与标准参考数值的误差大小为0.65%,运行时间为0.16秒;对于导体个数为8个的例子,本发明提取的寄生电容的数值结果与标准参考数值的误差大小为0.52%,运行时间为0.52秒;对于导体个数为71个的例子,本发明提取的寄生电容的数值结果与标准参考数值的误差大小为1.63%,运行时间为2.3秒;对于导体个数为97个的例子,本发明提取的寄生电容的数值结果与标准参考数值的误差大小为2.40%,运行时间为3.1秒。可以看到,对于个数越少的导体,本发明提取的寄生电容的误差越小,但对于个数庞大的导体,本发明提取的寄生电容的误差仍然不大,均控制在2.5%以内,且运行速度十分得快。

Claims (10)

  1. 基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,
    根据导体分布情况对集成电路版图进行非均匀网格划分,得到靠近导体边界均匀且稠密,远离导体边界不均匀且稀疏的网格形态;
    判断每个网格单元是否为边界单元,并对每个网格单元进行全局编号,对非边界单元的网格单元赋予求解单元编号,建立全局编号和求解单元编号的映射关系,非边界单元的网格单元组成需求解区域,所述边界单元为落在需求解区域边界外的网格单元以及导体所在区域内的网格单元;
    初始化边界单元的电势自由度,根据需求解区域内每个网格单元的相邻网格单元是否为边界单元的情形对局部间断有限元格式的静电场强公式简化,得到形如
    Figure PCTCN2022095540-appb-100001
    的线性方程组,求解所述线性方程组得到所有需求解网格单元电势自由度组合而成的电势函数自由度向量
    Figure PCTCN2022095540-appb-100002
    A为稀疏矩阵,
    Figure PCTCN2022095540-appb-100003
    为简化局部间断有限元格式的静电场强公式过程中产生的常数向量;
    依据局部间断有限元格式的静电场强公式以及所有需求解网格单元电势自由度组合而成的电势函数自由度向量,求解所有需求解网格单元电势自由度组合而成的电场强函数自由度向量;
    根据导体分布情况选取高斯面,在高斯面上进行电场强度函数的积分,获取导体电荷量,依据导体电荷量与电容的关系获取导体的电容以及导体互连线之间的寄生电容。
  2. 根据权利要求1所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据导体分布情况对集成电路版图进行非均匀网格划分的具体方法为:将每个导体划分为至少一个长方体,读取每个长方体的行列坐标,以两个相邻导体的边界行为基准,以一个极小值mingap的间隔距离均匀地向外扩展设定的次数,以gaptime倍数的mingap逐步扩大间隔距离,继续向外扩展直至扩展后的相邻两个导体的边界行之间的距离小于再次扩大后的间隔距离。
  3. 根据权利要求1所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,初始化边界单元的电势自由度的具体方法为:
    当边界单元不是主导体边界单元时,将该边界单元的电势自由度的常数值项赋值为0;
    当边界单元是主导体边界单元时,将该边界单元电势自由度的常数值项赋值为1,将其余边界单元电势自由度的常数值项赋值为0。
  4. 根据权利要求1所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据需求解区域内每个网格单元的相邻网格单元是否为边界单元的情形对局部间断有限元格式的静电场强公式简化的具体方法为:
    将所述局部间断有限元格式的静电场公式分为六个模块,
    模块一为:
    Figure PCTCN2022095540-appb-100004
    模块二为:
    Figure PCTCN2022095540-appb-100005
    模块三为:
    Figure PCTCN2022095540-appb-100006
    模块四为:
    Figure PCTCN2022095540-appb-100007
    模块五为:
    Figure PCTCN2022095540-appb-100008
    模块六为:
    Figure PCTCN2022095540-appb-100009
    其中,K xw,K xe表示网格单元k的左边界和右边界,K ys,K yn表示网格单元k的下边界和上边界,
    Figure PCTCN2022095540-appb-100010
    表示电场强q(x,-1/2)在网格单元k下边界的数值流通量大小,q(x,-1/2)表示网格单元k下边界处的电场强,
    Figure PCTCN2022095540-appb-100011
    表示电场强q(x,1/2)在网格单元k上边界处的数值流通量大小,q(x,1/2)表示网格单元k上边界处的电场强,
    Figure PCTCN2022095540-appb-100012
    表示电场强p(-1/2,y)在网格单元k左边界处的数值流通量大小,p(-1/2,y)表示网格单元k左边界处的电场强,
    Figure PCTCN2022095540-appb-100013
    表示电场强p(1/2,y)在网格单元k右边界处的数值流通量大小,p(1/2,y)表示网格单元k右边界处的电场强,
    Figure PCTCN2022095540-appb-100014
    表示基函数
    Figure PCTCN2022095540-appb-100015
    在网格单元k上边界的下方的流通量,
    Figure PCTCN2022095540-appb-100016
    表示网格单元k第l个基函数在上边界处的取值,
    Figure PCTCN2022095540-appb-100017
    表示基函数
    Figure PCTCN2022095540-appb-100018
    在网格单元k下边界的上方的流通量,
    Figure PCTCN2022095540-appb-100019
    表示网格单元k第l个基函数在下边界处的取值,
    Figure PCTCN2022095540-appb-100020
    表示基函数
    Figure PCTCN2022095540-appb-100021
    在网格单元k右边界的左方的流通量,
    Figure PCTCN2022095540-appb-100022
    表示网格单元k第l个基函数在右边界处的取值,
    Figure PCTCN2022095540-appb-100023
    表示基函数
    Figure PCTCN2022095540-appb-100024
    在网格单元k左边界的右方的流通量,
    Figure PCTCN2022095540-appb-100025
    表示网格单元k第l个基函数在左边界处的取值,p(x,y)表示网格单元k在x方向的电场强函数,q(x,y)表示网格单元k在y方向的电场强函数,
    Figure PCTCN2022095540-appb-100026
    表示网格单元k的第l个基函数;
    将六个模块的系数矩阵表示为11个矩阵的加减乘除运算的结果,所述11个 矩阵
    Figure PCTCN2022095540-appb-100027
    X nn、X ns、X sn、X ss、Y ee、Y ew、Y we、Y ww为表征参考网格单元上基函数与求解区域内各网格单元上基函数的转换关系的矩阵,
    Figure PCTCN2022095540-appb-100028
    Figure PCTCN2022095540-appb-100029
    Figure PCTCN2022095540-appb-100030
    Figure PCTCN2022095540-appb-100031
    Figure PCTCN2022095540-appb-100032
    Figure PCTCN2022095540-appb-100033
    Figure PCTCN2022095540-appb-100034
    Figure PCTCN2022095540-appb-100035
    Figure PCTCN2022095540-appb-100036
    Figure PCTCN2022095540-appb-100037
    Figure PCTCN2022095540-appb-100038
    其中,
    Figure PCTCN2022095540-appb-100039
    表示参考网格单元K base第i个基函数
    Figure PCTCN2022095540-appb-100040
    在第j个基函数
    Figure PCTCN2022095540-appb-100041
    上的投影,
    Figure PCTCN2022095540-appb-100042
    表示参考网格单元K base第i个基函数x方向的导数在第j个基函数
    Figure PCTCN2022095540-appb-100043
    上的投影,
    Figure PCTCN2022095540-appb-100044
    表示参考网格单元K base第i个基函数y方向的导数在第j个基函数
    Figure PCTCN2022095540-appb-100045
    上的投影,
    Figure PCTCN2022095540-appb-100046
    表示参考网格单元K base第j个基函数在下边界取值
    Figure PCTCN2022095540-appb-100047
    与第i个基函数在下边界取值
    Figure PCTCN2022095540-appb-100048
    的乘积,
    Figure PCTCN2022095540-appb-100049
    表示参考网格单元K base第j个基函数在上边界取值
    Figure PCTCN2022095540-appb-100050
    与第i个基函数在下边界取值
    Figure PCTCN2022095540-appb-100051
    的乘积,
    Figure PCTCN2022095540-appb-100052
    表示参考网格单元K base第j个基函数在下边界取值
    Figure PCTCN2022095540-appb-100053
    与第i个基函数在上边界取值
    Figure PCTCN2022095540-appb-100054
    的乘积,
    Figure PCTCN2022095540-appb-100055
    表示参考网格单元K base第j个基函数在上边界取值
    Figure PCTCN2022095540-appb-100056
    与第i个基函数在上边界取值
    Figure PCTCN2022095540-appb-100057
    Figure PCTCN2022095540-appb-100058
    的乘积,
    Figure PCTCN2022095540-appb-100059
    表示参考网格单元K base第j个基函数在左边界取值
    Figure PCTCN2022095540-appb-100060
    与第i个基函数在左边界取值
    Figure PCTCN2022095540-appb-100061
    的乘积,
    Figure PCTCN2022095540-appb-100062
    表示参考网格单元K base第j个基函数在右边界取值
    Figure PCTCN2022095540-appb-100063
    与第i个基函数在左边界取值
    Figure PCTCN2022095540-appb-100064
    的乘 积,
    Figure PCTCN2022095540-appb-100065
    表示参考网格单元K base第j个基函数在左边界的取值
    Figure PCTCN2022095540-appb-100066
    与第i个基函数在右边界取值
    Figure PCTCN2022095540-appb-100067
    的乘积,
    Figure PCTCN2022095540-appb-100068
    表示参考网格单元K base第j个基函数在右边界取值
    Figure PCTCN2022095540-appb-100069
    与第i个基函数在右边界取值
    Figure PCTCN2022095540-appb-100070
    的乘积;
    遍历需求解区域内的每个网格单元,根据与当前网格单元的上边界相邻的网格单元、以及与上边界网格单元的上边界相邻的网格单元是否为边界单元的情形对模块一进行简化,根据与当前网格单元的上边界相邻的网格单元是否为边界单元的情形对模块二进行简化,根据与当前网格单元的上边界相邻的网格单元是否为边界单元的情形对模块三进行简化,根据与当前网格单元的右边界相邻的网格单元、以及与右边界网格单元的右边界相邻的网格单元是否为边界单元的情形对模块四进行简化,根据与当前网格单元的右边界相邻的网格单元是否为边界单元的情形对模块五进行简化,根据与当前网格单元的右边界相邻的网格单元是否为边界单元的情形对模块六进行简化。
  5. 根据权利要求4所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的上边界相邻的网格单元、以及与上边界网格单元的上边界相邻的网格单元是否为边界单元的情形对模块一进行简化的具体方法为:
    与当前网格单元的上边界相邻的网格单元、以及与上边界网格单元的上边界相邻的网格单元均为非边界单元时,模块一简化为:
    Figure PCTCN2022095540-appb-100071
    与当前网格单元的上边界相邻的网格单元为非边界单元,与上边界网格单元的上边界相邻的网格单元为边界单元为边界单元时,模块一简化为:
    Figure PCTCN2022095540-appb-100072
    与当前网格单元的上边界相邻的网格单元为边界单元时,模块一简化为:
    Figure PCTCN2022095540-appb-100073
    其中,Δx k表示网格单元k在x方向的长度,Δy k表示网格单元k在y方向上的长度,Δy kn表示网格单元k上方的邻近网格单元在y方向上的长度,
    Figure PCTCN2022095540-appb-100074
    表示网 格单元k电势函数自由度向量,
    Figure PCTCN2022095540-appb-100075
    表示与网格单元k上边界相邻的网格单元的电势函数自由度向量,
    Figure PCTCN2022095540-appb-100076
    表示与网格单元k下边界相邻的网格单元的电势函数自由度向量,
    Figure PCTCN2022095540-appb-100077
    表示与网格单元k上边界网格单元的上边界相邻的网格单元的电势函数自由度向量。
  6. 根据权利要求5所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的上边界相邻的网格单元是否为边界单元的情形对模块二进行简化的具体方法为:
    与当前网格单元的上边界相邻的网格单元为非边界单元时,模块二简化为:
    Figure PCTCN2022095540-appb-100078
    与当前网格单元的上边界相邻的网格单元为边界单元时,模块二简化为:
    Figure PCTCN2022095540-appb-100079
  7. 根据权利要求6所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的上边界相邻的网格单元是否为边界单元的情形对模块三进行简化的具体方法为:
    与当前网格单元的上边界相邻的网格单元为非边界单元时,模块三简化为:
    Figure PCTCN2022095540-appb-100080
    与当前网格单元的上边界相邻的网格单元为边界单元时,模块三简化为:
    Figure PCTCN2022095540-appb-100081
  8. 根据权利要求7所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的右边界相邻的网格单元、以及与右边界网格单元的右边界相邻的网格单元是否为边界单元的情形对模块四进行简化的具体方法为:
    与当前网格单元的右边界相邻的网格单元、以及与右边界网格单元的右边界相邻的网格单元均为非边界单元时,模块四简化为:
    Figure PCTCN2022095540-appb-100082
    与当前网格单元的右边界相邻的网格单元为非边界单元,与右边界网格单元的右边界相邻的网格单元为边界单元时,模块四简化为:
    Figure PCTCN2022095540-appb-100083
    与当前网格单元的右边界相邻的网格单元为边界单元时,模块四简化为:
    Figure PCTCN2022095540-appb-100084
    其中,Δx ke表示网格单元k右方的邻近网格单元在x方向上的长度,
    Figure PCTCN2022095540-appb-100085
    表示与网格单元k的右边界相邻的网格单元的电势函数自由度向量,
    Figure PCTCN2022095540-appb-100086
    表示与网格单元k右边界网格单元的右边界相邻的网格单元的电势函数自由度向量,
    Figure PCTCN2022095540-appb-100087
    表示与网格单元k左边界相邻的网格单元的电势函数自由度向量。
  9. 根据权利要求8所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的右边界相邻的网格单元是否为边界单元的情形对模块五进行简化的具体方法为:
    与当前网格单元的右边界相邻的网格单元为非边界单元时,模块五简化为:
    Figure PCTCN2022095540-appb-100088
    与当前网格单元的右边界相邻的网格单元为边界单元时,模块五简化为:
    Figure PCTCN2022095540-appb-100089
    其中,
    Figure PCTCN2022095540-appb-100090
    表示与网格单元k右边界相邻的网格单元的电势函数自由度向量。
  10. 根据权利要求9所述基于间断有限元法的集成电路互连线寄生电容提取方法,其特征在于,根据与当前网格单元的右边界相邻网格单元是否为边界单元的情形对模块六进行简化的具体方法为:
    与当前网格单元的右边界相邻的网格单元为非边界单元时,模块六简化为:
    Figure PCTCN2022095540-appb-100091
    与当前网格单元的右边界相邻的网格单元为边界单元时,模块六简化为:
    Figure PCTCN2022095540-appb-100092
PCT/CN2022/095540 2022-03-17 2022-05-27 基于间断有限元法的集成电路互连线寄生电容提取方法 WO2023173592A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210263251.3A CN114357942B (zh) 2022-03-17 2022-03-17 基于间断有限元法的集成电路互连线寄生电容提取方法
CN202210263251.3 2022-03-17

Publications (1)

Publication Number Publication Date
WO2023173592A1 true WO2023173592A1 (zh) 2023-09-21

Family

ID=81094359

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/095540 WO2023173592A1 (zh) 2022-03-17 2022-05-27 基于间断有限元法的集成电路互连线寄生电容提取方法

Country Status (2)

Country Link
CN (1) CN114357942B (zh)
WO (1) WO2023173592A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117371385A (zh) * 2023-12-06 2024-01-09 杭州行芯科技有限公司 邻近图形的确认方法、电子设备及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114357942B (zh) * 2022-03-17 2022-06-10 南京邮电大学 基于间断有限元法的集成电路互连线寄生电容提取方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120074233A1 (en) * 2010-08-12 2012-03-29 Féinics AmaTech Nominee Limited Coupling in and to rfid smart cards
CN103164572A (zh) * 2013-02-22 2013-06-19 中国科学院电工研究所 一种集成电路互连线寄生电容的建模方法
EP2858246A1 (en) * 2013-10-07 2015-04-08 Dialog Semiconductor GmbH Capacitance minimization switch
CN105895145A (zh) * 2014-05-09 2016-08-24 瑞萨电子株式会社 半导体器件
CN114357942A (zh) * 2022-03-17 2022-04-15 南京邮电大学 基于间断有限元法的集成电路互连线寄生电容提取方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2808101A1 (fr) * 1999-08-31 2001-10-26 Commissariat Energie Atomique Calcul de capacites entre conducteurs dans un circuit a etages multiples
CN1881562A (zh) * 2005-06-13 2006-12-20 北京中电华大电子设计有限责任公司 一种自动适应工艺特征尺寸的互连寄生电容提取方法
US8219377B2 (en) * 2009-02-23 2012-07-10 Georgia Tech Research Corporation Multi-layer finite element method for modeling of package power and ground planes
CN102136449B (zh) * 2010-01-22 2013-11-27 复旦大学 基于嵌套式准二阶随机配置法的寄生电容提取方法
CN102521471B (zh) * 2012-01-02 2013-09-25 西安电子科技大学 基于边界元法的矩形冗余填充耦合电容提取方法
CN110852508A (zh) * 2019-11-11 2020-02-28 吉林大学 一种多导体传输线串扰频域概率分布的预测方法
CN110988993B (zh) * 2019-11-27 2021-01-26 清华大学 一种偏移成像方法、装置及电子设备
CN112466769B (zh) * 2020-11-20 2024-01-30 西安电子科技大学 圆形电容结构的hemt器件沟道区电场分布测量结构和方法
CN112487755B (zh) * 2020-12-15 2022-12-09 西安交通大学 一种fltd腔体中瞬态电磁场分布的数值计算方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120074233A1 (en) * 2010-08-12 2012-03-29 Féinics AmaTech Nominee Limited Coupling in and to rfid smart cards
CN103164572A (zh) * 2013-02-22 2013-06-19 中国科学院电工研究所 一种集成电路互连线寄生电容的建模方法
EP2858246A1 (en) * 2013-10-07 2015-04-08 Dialog Semiconductor GmbH Capacitance minimization switch
CN105895145A (zh) * 2014-05-09 2016-08-24 瑞萨电子株式会社 半导体器件
CN114357942A (zh) * 2022-03-17 2022-04-15 南京邮电大学 基于间断有限元法的集成电路互连线寄生电容提取方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117371385A (zh) * 2023-12-06 2024-01-09 杭州行芯科技有限公司 邻近图形的确认方法、电子设备及存储介质
CN117371385B (zh) * 2023-12-06 2024-04-09 杭州行芯科技有限公司 邻近图形的确认方法、电子设备及存储介质

Also Published As

Publication number Publication date
CN114357942A (zh) 2022-04-15
CN114357942B (zh) 2022-06-10

Similar Documents

Publication Publication Date Title
WO2023173592A1 (zh) 基于间断有限元法的集成电路互连线寄生电容提取方法
US10169506B2 (en) Circuit design method and system
TWI709053B (zh) 積體電路設計系統以及用於積體電路設計之方法
CN103544333B (zh) 半导体器件设计方法、系统和计算机程序产品
US8887116B2 (en) Flexible pattern-oriented 3D profile for advanced process nodes
US20060156266A1 (en) Probabilistic congestion prediction with partial blockages
TWI521370B (zh) 用於產生時鐘網路拓撲之暫存器叢集方法、系統及其電腦可讀取記錄媒體
US10083266B2 (en) Simulation method of CMP process
CN102364480B (zh) 提取寄生参数的方法及系统
TWI575394B (zh) 透過考量不同電路拓撲之輸入波形產生來特徵化元件
Abouelyazid et al. Connectivity-based machine learning compact models for interconnect parasitic capacitances
KR101495986B1 (ko) 더블 패터닝 기술을 위한 rc 코너해
TWI817646B (zh) 用於設計上下文感知電路的方法及系統
CN106094422A (zh) 一种简化opc后掩模版图形的方法
CN106815379B (zh) 一种提取寄生电容的方法及系统
US6542834B1 (en) Capacitance estimation
CN109841532A (zh) 制造半导体装置的方法
CN114611455B (zh) 一种基于改进有限差分法的二维电容提取方法
CN104600066B (zh) 定义氧化层(od)梯度减小的半导体器件及其制作方法
TWI712904B (zh) 電子裝置及積體電路的佈局方法
US7739639B2 (en) Method and apparatus of core timing prediction of core logic in the chip-level implementation process through an over-core window on a chip-level routing layer
Wang et al. Short critical area model and extraction algorithm based on defect characteristics in integrated circuits
CN112214955B (zh) 一种提取超大规模集成电路芯片电源模型参数的方法
US11328873B2 (en) Parallel plate capacitor resistance modeling and extraction
Patidar et al. Evaluation Trends and Development In Integrated Circuit Parasitic Extraction

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 18022272

Country of ref document: US

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22931617

Country of ref document: EP

Kind code of ref document: A1