Method for extracting and storing DC resistance model, computer device and storage medium
By constructing virtual rule Octet tree node unit and feature tree merging, the complexity and storage problems of Octet tree structure in large-scale data processing are solved, the accuracy and adaptability of boundary processing are improved, and efficient resistance model storage is achieved. and search.
Patent Information
- Application Number
- CN202411306465.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-19
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-09-19
AI Technical Summary
The existing octree structure has high algorithm complexity, large storage space requirements, insufficient boundary processing and poor adaptability when processing large-scale data, resulting in high computational costs, large errors, large storage resource pressure and inability to adapt to complex data characteristics.
By constructing a virtual rule octree node unit, simulate and divide the irregular units, calculate the resistance value using the relationship between the regular units and the preset function, reduce the data storage amount, adopt a linear index storage method, merge the feature tree of the same feature, and convert it into Quadtree structure reduces the amount of calculation.
It improves the accuracy of boundary division, reduces calculation complexity and time consumption, reduces storage space requirements, improves processing efficiency and adaptability, reduces boundary effects, and achieves more accurate resistance model storage and retrieval.
Smart Images

Figure CN118821714B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of resistance model analysis of system-level chips, and in particular to a method for extracting and storing a direct current resistance model, and a computer device and a computer-readable storage medium for implementing the method. Background Art
[0002] With the development of integrated circuit technology, three-dimensional integrated circuits are widely used in large-capacity storage and high-speed signal processors due to their advantages such as high density, high speed, multi-function and low power consumption. Since three-dimensional integrated circuits have a large number of components, it is necessary to extract and analyze the resistance model of the three-dimensional integrated circuit. The existing resistance model extraction and storage method is based on the octree structure. For example, the invention patent application with publication number CN114529648A discloses a model display method, which is based on the octree structure model.
[0003] As an important three-dimensional spatial data structure, the octree structure plays a key role in many fields, especially in the processing of complex three-dimensional data and scenes. The octree structure recursively divides the three-dimensional space into eight equal subspaces, each of which is a sub-cube, providing a solid data structure foundation for the storage, retrieval, analysis and visualization of three-dimensional data. The octree structure is used in many key areas of computer science, including three-dimensional modeling, point cloud processing, computer graphics, geographic information systems (GIS), and electromagnetic simulation in finite element meshing technology.
[0004] Although the octree structure technology has many advantages, it still has the following shortcomings in practical applications:
[0005] First, the algorithm complexity of the existing octree structure is high. The construction and query algorithms of the octree are relatively complex. Especially when processing large-scale data, a large number of recursive operations are required, which may lead to high computational costs and time consumption.
[0006] Second, the existing octree structure consumes a lot of storage space. As the data size increases, the storage space required by the octree structure will also increase significantly, which may put pressure on the system's storage resources.
[0007] Third, the existing octree structure has deficiencies in boundary processing. When processing shapes with complex boundaries, the existing octree structure often cannot accurately represent the boundaries, resulting in boundary effects and errors in data calculation.
[0008] Fourth, the existing octree structure has poor adaptability. In some specific application scenarios, the division method of the octree structure is not flexible enough and it is difficult to adapt to complex and changeable data characteristics. Summary of the invention
[0009] The first object of the present invention is to provide a method for extracting and storing a DC resistance model that can improve the accuracy of boundary division.
[0010] The second object of the present invention is to provide a computer device for implementing the above-mentioned DC resistance model extraction and storage method.
[0011] The third object of the present invention is to provide a storage medium for implementing the above-mentioned DC resistance model extraction and storage method.
[0012] To achieve the first objective of the present invention, the method for extracting and storing a DC resistance model provided by the present invention includes obtaining layout information of an integrated circuit to be analyzed, dividing the elements of the power network of the integrated circuit into an octree structure according to the layout information; calculating the minimum division matrix unit of each power network, and dividing each child node of each power network into an octree structure, and mapping and storing the data of each power network into a multi-level storage system according to the division result of the octree structure; calculating the resistance value of each node in the octree structure according to the conductivity of different material regions in the integrated circuit, and storing the resistance value of each node; after receiving a display instruction, displaying the set The invention discloses an octree structure for forming a circuit and the resistance value of each node of the octree structure; wherein, the octree structure division for each sub-node of each power supply network includes: dividing the octree structure for each sub-node, judging whether the unit of the current first-level octree structure is divided into the minimum division matrix unit, and if not, continuing to divide the unit of the current first-level octree structure into the octree structure until the unit of the current first-level octree structure is divided into the minimum division matrix unit; and, for the irregular unit on the octree structure, constructing a virtual regular octree node unit by using the regular unit and the preset function relationship, and calculating the resistance value of the virtual regular octree node unit.
[0013] It can be seen from the above scheme that for irregular units at the boundaries of the integrated circuit, the irregular units are simulated into regular units by constructing virtual regular octree node units, so that each node on the octree structure is of regular shape, which is convenient for data storage of the octree structure and data retrieval of the octree structure. Since there are often curved surfaces at irregular units, in order to accurately calculate the resistance value of each octree node, the resistance value of the virtual regular octree node unit is calculated by using regular units and preset function relationships, so as to accurately reflect the resistance value of the virtual regular octree node unit, so that the resistance value calculation of the integrated circuit at the boundary is more accurate and the boundary effect can be reduced.
[0014] A preferred solution is that the regular unit is a regular cubic unit or a cuboid with square upper and lower surfaces, and the irregular unit is a unit other than a regular cubic unit or a cuboid with square upper and lower surfaces.
[0015] It can be seen that all shapes other than regular cubes are defined as irregular units, and when calculating the resistance value of the irregular unit, a regular octree structure is formed by simulation using a regular cube, which can facilitate data storage and retrieval of the octree structure.
[0016] A further solution is that the octree structure division of each subnode of each power supply network also includes: after the octree structure division of a target subnode, the characteristics of the target subnode are recorded and a feature tree is formed, and when the octree structure division of subsequent subnodes is performed, if the subsequent subnodes have the same characteristics as the target subnode, the octree structure of the target subnode is directly used as the octree structure of the subsequent subnode, and the coordinate offset between the subsequent subnode and the target subnode is recorded.
[0017] Since different elements in the octree structure often have similar features, in order to save the data storage capacity of the octree structure, for the data storage of subsequent child nodes and the target child node with the same features, the octree structure of the target child node is directly used as the octree structure of the subsequent child nodes, avoiding the storage of a large amount of similar data, thereby reducing the data storage capacity of the octree structure.
[0018] A further solution is that determining that the subsequent subnode has the same feature as the target subnode includes: the subsequent subnode has the same via radius and / or uses the same material as the target subnode.
[0019] A further solution is that the octree structure division of each subnode of each power network also includes: if the power network is a device layer power network, obtaining a first feature tree of the first type of element and a second feature tree of the second type of element, and then merging the first feature tree and the second feature tree according to the overlapping nodes of the first type of element and the second type of element to form a third feature tree.
[0020] It can be seen that since there are often a large number of overlapping elements of different types in the device layer of the power supply network, the feature trees of different types of elements are extracted for the overlapping areas, and the first feature tree is merged with the second feature tree to form a third feature tree, thereby forming a feature tree formed for the overlap of two types of elements.
[0021] A further solution is to record the number of the first feature tree and the number of the second feature tree when storing the data of the third feature tree, and calculate the merging function of the first feature tree and the second feature tree to form a directional general feature tree.
[0022] It can be seen that since the third feature tree is formed based on the first feature tree and the second feature tree, in order to reduce the amount of data storage, the data storage of the third feature tree only records the numbers of the first feature tree and the second feature tree, and calculates the merge function of the first feature tree and the second feature tree, which can further reduce the data storage capacity of the octree structure.
[0023] A further solution is that when calculating the minimum partitioning matrix unit of each power network, the length of the minimum partitioning matrix unit is determined by the minimum length of each element of the power network, the width of the minimum partitioning matrix unit is determined by the minimum width of each element of the power network, and the height of the minimum partitioning matrix unit is set to a preset value.
[0024] Since the layer height of an integrated circuit is often fixed, when calculating the size of the minimum partitioning matrix unit, the length and width of the minimum partitioning matrix unit are calculated, and the height is set to a preset value, which can convert the octree structure into a quadtree structure and reduce the amount of calculation for octree partitioning.
[0025] A further solution is to map and store the data of each power supply network into a multi-level storage system, including: mapping each node of the octree structure of the power supply network into a one-dimensional array in a linear index manner, and determining the position of each node in the array according to the position of the parent node of each node and the order of the child nodes of the node in the parent node.
[0026] It can be seen that mapping each node of the octree structure of the power supply network into a one-dimensional array by means of linear indexing can standardize the data storage operation of the octree structure, facilitate the subsequent data retrieval of the octree structure, and also facilitate the data reading operation of the octree structure.
[0027] To achieve the above-mentioned second purpose, the computer device provided by the present invention includes a processor and a memory, and the memory stores a computer program. When the computer program is executed by the processor, each step of the above-mentioned DC resistance model extraction and storage method is implemented.
[0028] To achieve the third objective mentioned above, a computer program is stored on the storage medium provided by the present invention, and when the computer program is executed by a processor, each step of the above-mentioned method for extracting and storing the DC resistance model is implemented. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a flow chart of an embodiment of the method for extracting and storing a DC resistance model of the present invention.
[0030] Figure 2 It is a data storage schematic diagram of an octree structure used in an embodiment of a method for extracting and storing a DC resistance model of the present invention.
[0031] Figure 3 It is a schematic diagram of the structure of a multi-level storage system established in the embodiment of the method for extracting and storing a DC resistance model of the present invention.
[0032] Figure 4 It is a schematic diagram of the octree structure used in the embodiment of the method for extracting and storing the DC resistance model of the present invention.
[0033] Figure 5 It is a schematic diagram of chip BGA topology.
[0034] Figure 6 It is a schematic diagram of the octree structure of the first feature tree extracted by the embodiment of the method for extracting and storing a DC resistance model of the present invention.
[0035] Figure 7 It is a schematic diagram of the octree structure of the second feature tree extracted by the embodiment of the method for extracting and storing a DC resistance model of the present invention.
[0036] Figure 8 It is a schematic diagram of the structure of the third feature tree formed by the embodiment of the method for extracting and storing a DC resistance model of the present invention.
[0037] The present invention is further described below in conjunction with the accompanying drawings and embodiments. DETAILED DESCRIPTION
[0038] The present invention is directed to data storage of a resistance model of an integrated circuit. Specifically, the resistance model of the integrated circuit is stored based on an octree structure. The present invention improves the accuracy of data calculation at the boundary by constructing a virtual regular octree node unit. The computer device of the present invention has a processor and a memory, the memory is a readable storage medium of the present invention, a computer program is stored on the memory, and when the computer program is executed by the processor, the above-mentioned method for extracting and storing the DC resistance model is implemented.
[0039] DC resistance model extraction and storage method embodiment:
[0040] See also Figure 1 In this embodiment, step S11 is first performed to obtain the layout information of the three-dimensional integrated circuit to be analyzed, such as reading the layout file of the three-dimensional integrated circuit, and calculating the overall size information of the layout, including the length, width, height and other information of the layout. Then, based on the read layout information, the layout of the integrated circuit is divided into power supply networks, including identifying pads, vias, copper shapes, connections and other elements in the layout, sorting them by position and classifying them into corresponding power supply network data, and numbering and marking each power supply network, for example, marking each power supply network as NETi, where i is 0, 1, 2..., which is the serial number of the power supply network.
[0041] Next, an N-ary tree structure of each power network NETi is established, and the data of the N-ary tree structure of each power network NETi is marked as NETi_Tree_data. The root node of the N-ary tree structure is the power network number, that is, the ID of the power network. Each PIN pin of the N-ary tree structure of each power network NETi is a secondary child node, such as Figure 2 As shown, each power network NETi has a root node 10, which is divided into multiple secondary child nodes, such as child nodes 11, 12, 13, etc. If the PIN pin in a certain direction has no coordinates, the child node in this direction is empty. Therefore, the N-ary tree structure divided under the root node 10 does not necessarily satisfy that all eight directions have secondary child nodes, so it becomes an N-ary tree structure.
[0042] For each N-ary tree structure of the power network NETi, the corresponding data needs to be stored in a multi-level storage system. The storage structure of the multi-level storage system is as follows: Figure 3As shown, the multi-level storage system is provided with a power network ID table 21, a network and congruent matrix unit mapping table 22, a unit matrix storage structure 23 and a resistance value table 24. Among them, the power network ID table 21 records the ID of each power network, such as the power network M1ID1, etc. The network and congruent matrix unit mapping table 22 records the mapping relationship between each power network and the congruent matrix unit. For example, if a certain power network corresponds to one or more congruent matrix units, it is necessary to record the ID of each congruent matrix unit corresponding to the power network. The unit matrix storage structure 23 records the resistance value ID of each congruent matrix unit in different dimensions, wherein the "dimension" in the "value" of the unit matrix storage structure 23 is the direction of the three-dimensional integrated circuit, such as along the positive direction of the X-axis, the negative direction of the X-axis or the positive direction of the Y-axis, and the negative direction of the Y-axis. The resistance value table 24 records the specific values of the resistance values corresponding to each resistance value ID. Among them, the congruent matrix unit can be the minimum partition matrix unit introduced later, or it can be a unit of an integer multiple of the minimum partition matrix unit.
[0043] Of course, when establishing the storage structure of the multi-level storage system, since the resistance value of each unit node of each power network NETi is not calculated, the specific resistance value of each congruent matrix unit is not filled.
[0044] In addition, it is necessary to model the layout of the integrated circuit, modify the stacking structure, select parameter information such as the load, material and input power information of the power network, and then start the extraction operation of the DC resistance model.
[0045] Next, execute step S12 to calculate the minimum partition matrix unit of each power network. For example, a subnode unit of the selected power network is marked as PinNode-n, and the minimum partition matrix unit of each power network NETi is marked as Matrix_Min_i. The calculation method of the minimum partition matrix unit is as follows: First, record the length L and width W of each element of the power network NETi, and find out the minimum length L and minimum width W of each element, and determine the size of the minimum partition matrix unit based on the minimum length L and width W of each element or the integer multiple of the minimum length L and width W. The minimum length L and minimum width W of the minimum partition matrix unit are calculated by the following formula:
[0046] length Lmin = MIN([number1]length, [number2]length, ...);
[0047] Width Wmin = MIN([number1]width, [number2]width, ...).
[0048] Among them, number1 and number2 are various elements in the power network, such as wiring, vias and other conductive areas.
[0049] Therefore, if the length Lmin is greater than the width Wmin, the size of the minimum partition matrix unit = width Wmin×width Wmin×unit height / 2; otherwise, the size of the minimum partition matrix unit = length Lmin×length Lmin×unit height / 2.
[0050] Since the three-dimensional integrated circuit is a stacked structure, the height H of each stacked structure is usually uniform. Therefore, when calculating the size of the minimum partition matrix unit, the height of the minimum partition matrix unit can be directly set to a preset value.
[0051] Preferably, the minimum partitioning matrix unit Matrix_Min_i can enlarge or reduce the length and width by integer multiples according to the computing resources. For example, for a cylinder with a diameter of 1 nanometer, assuming that the layer height is 1 nanometer, the calculated minimum partitioning matrix unit Matrix_Min_i is a regular hexahedron with a width W of 0.5 nanometers, a length L of 0.5 nanometers, and a height H of 0.25 nanometers. The aforementioned integer multiple refinement of the unit hexahedron is to divide the length L and the width W by an integer multiple n, and the height H remains unchanged. For example, the width W of the 2-fold refined unit hexahedron is 0.25 nanometers, the length L is 0.25 nanometers, and the height H is 0.25 nanometers. It is very necessary to calculate the size of the minimum partitioning matrix unit Matrix_Min_i, because the control of granularity is crucial to the feasibility and optimization of the subsequent power network octree structure division. The extraction of the minimum partition matrix unit Matrix_Min_i is similar to setting the resolution reasonably during 3D modeling. The depth of the partitioned octree structure is reduced according to this threshold, while the complex graphics of the integrated circuit layout are reasonably divided, such as the narrow edges brought by circles, nano-sized wires, etc. Complex polygons can be simplified by polygons to obtain equally divided hexahedrons.
[0052] Then, step S13 is executed to divide each child node of the power network into an octree structure. Preferably, since there are a large number of child nodes of the power network, multiple threads of the thread queue or multiple processes of the process queue can be started to divide the multiple child nodes into an octree structure in parallel, and a processing business process or thread is marked as PThread_i.
[0053] The partitioning process of the octree structure is as follows Figure 4As shown, a region is divided into 8 cells, that is, eight tree nodes. If a cell does not meet the conditions for stopping recursive division, the next level of recursive division needs to be performed until the conditions are met. Assuming that a child node 31 is a regular cube, when performing octree division on the child node 31, the child node 31 is divided into eight equally sized refined cubes to form a cell 32 of the octree structure of the next level of refinement, and each cell 32 can also be further refined and divided again into eight equally sized refined cubes to form a cell 33 of the octree structure of the next level, and so on. Combined with Figure 2 Preferably, each secondary child node can be further divided into multiple subtree root nodes, for example, subtree root node 14 is a lower-level child node of child node 11, subtree root node 15 is a lower-level child node of child node 13, and so on. In addition, for each subtree root node, the division is continued according to the octree structure, for example, for each subtree root node, it is divided into cubes of equal size according to its eight directions, and each cube is numbered, for example, 0 represents the unit in the top upper left corner, 1 represents the unit in the top upper right corner, 2 represents the unit in the top lower left corner, 3 represents the unit in the top lower right corner, 4 represents the unit in the bottom upper left corner, 5 represents the unit in the bottom upper right corner, 6 represents the unit in the bottom lower left corner, and 7 represents the unit in the bottom lower right corner. In addition, when storing the data of the octree structure in a multi-level storage system, it is stored in a linear index storage manner. Specifically, the nodes of the octree structure are mapped to a one-dimensional array, and the position of each node in the one-dimensional array is determined by the position of its parent node and the order of the node in its parent node child nodes. This indexing method usually requires a clear rule to calculate the corresponding index value from the three-dimensional coordinates, or to infer the three-dimensional coordinates from the index value. Since each stacked structure is basically uniform, the bottom node unit of the octree of each element can be calculated based on the top node unit. Therefore, when storing data, only the data of the top four units of the octree structure need to be stored, and the calculation relationship of the bottom four units need to be set, without storing the data of eight units. The data of the octree structure is converted into the data of the quadtree structure, thereby reducing the data storage space.
[0054] The specific steps of dividing the child node PinNode-n on a power network NETi into an octree structure are as follows:
[0055] The first step is to establish an octree structure for a certain element. The root node of the octree structure is identified by the element name and ID. Each element of the power network NETi is read out in order and according to the position of the PIN pin. It is determined whether the current element needs to be simplified. If it needs to be simplified, polygon simplification is performed and then the second step is executed. Otherwise, the second step is executed directly.
[0056] The second step is to divide the current unit into an octree structure, and recursively check whether each unit meets the size of the minimum division matrix unit Matrix_Min_i. If it meets the size of the minimum division matrix unit Matrix_Min_i, stop dividing the element. If not, continue to recursively divide the unit into an octree structure.
[0057] The third step is to store the center coordinates of each unit divided in each direction in sequence according to the final result of the octree structure division of each element of the unit network NETi, and store the divided units in sequence in a certain direction to form an octree node.
[0058] The fourth step is that since the elements of each child node PinNode-n often have similar features, for example, the radius of each via is the same, or the filled material is the same, but the name and coordinates of each child node PinNode-n are different, for child nodes with the same features, after the first octree structure division for such child nodes, this type of child node is marked as a feature tree, and the child node of the first octree structure division is called the target child node. When the octree structure is subsequently divided for the same child node, if the subsequent child node has the same features as the target child node, the octree structure of the target child node is directly used as the octree structure of the subsequent child node, and the coordinate offset between the subsequent child node and the target child node is recorded. This can further reduce the amount of data storage in the octree structure.
[0059] In the fifth step, after dividing the octree structure for multiple elements, Boolean operations are performed on the octree structures formed by each element, and they are merged into the node branch of the child node PinNode-n on the power network NETi to form the octree structure of the child node PinNode-n on the power network NETi.
[0060] It should be noted that, since the octree structure data NETi_Tree_data of the power network is the structure data of an N-ary tree, the octree structure data of the child node PinNode-n is generated for each child node of the N-ary tree structure data. The characteristics of the child node can be identified according to the deep search algorithm, and the feature tree can be extracted from it. Then, the direction of the feature tree and other information are extracted for compression and storage, and finally mapped to the child nodes of the N-ary tree of the octree structure data NETi_Tree_data of the power network. In addition, for the data of each feature tree, a decision tree can be constructed using the feature tree by machine learning, and the data set can be divided into feature trees of different categories by feature partitioning again, providing necessary data for the subsequent automatic intelligent extraction of feature trees and classification of feature trees.
[0061] Since the integrated circuit has a large number of irregular units, for example, the boundary of a cylindrical via is not a regular cube, but a structure with a curved surface. In this embodiment, a regular cube unit or a rectangular parallelepiped with a square upper and lower surface is called a regular unit. If a unit is not a regular cube structure, for example, it has a curved surface, it is called an irregular unit, that is, an irregular unit is a unit outside the regular cube structure. An irregular unit may have a part of a metal conductor and another part of an insulating medium. When the octree structure is divided, the division of these irregular units will bring a large amount of irregular data. In order to facilitate data storage, this embodiment uses regular units and preset functional relationships to construct virtual regular octree node units, and calculates the resistance value of the virtual regular octree node unit. For example, the preset regular unit is a metal copper with a length, width and height of 1 nanometer. If a certain irregular unit has a curved surface, half of it is copper and the other half is an insulating medium. At this time, the resistance value corresponding to the irregular unit can be calculated according to a preset functional relationship. For example, the functional relationship can be a linear equation matching the curved surface.
[0062] In addition, if a power supply network is a device layer power supply network, when dividing the octree structure, the feature tree can be fully utilized for combination. Figure 5 As shown in the diagram of an integrated circuit, the red area is the power network 42, the green area is the ground network 41, a large number of vias 43 and pads 44 are formed on the integrated circuit, and there is a non-metallic area 45, and the number 46 is the center of a certain element. Assume that the device layer has a uniform thickness and a height of H.
[0063] For the octree structure division of the device layer power network, for example, for the power network, after obtaining a certain element, it is necessary to first query the feature tree data to determine whether there is a corresponding feature tree for the element that needs to be divided, that is, whether there is a feature tree that matches the current element, such as whether there are vias with the same aperture and filling material.
[0064] Then, feature trees are extracted according to different types of elements. Specifically, the octree structure of pad 44 is divided as follows: Figure 6 As shown, the pad 44 constitutes the first type of element of this embodiment, Figure 6 (a), (b), (c), and (d) are respectively the top view, bottom view, front view, and rear view of the pad 44 after the octree structure is divided. It can be seen that after the pad 44 is divided into the octree structure, the four top units are 51, 52, 53, and 54, and the four bottom units are 55, 56, 57, and 58. By extracting data of this structure, a first feature tree can be formed.
[0065] The octree structure division for the through line 43 is shown in FIG7 . The through line 43 constitutes the second type of element of this embodiment. Figure 7 (a), (b), (c), and (d) are respectively the top view, front view, rear view, and bottom view of the octree structure division of the line 43. It can be seen that after the line 43 is divided, 16 units will be formed, of which the top eight units are 61, 62, 63, 64, 65, 66, 67, and 68, and the bottom eight units are 71, 72, 73, 74, 75, 76, 77, and 78. By extracting data of this structure, a second feature tree can be formed.
[0066] Then, the first feature tree and the second feature tree are merged according to the overlapping nodes of the first type of elements and the second type of elements to form a third feature tree, such as Figure 8 As shown, assuming that the center point of the pad 44 coincides with the left midpoint of the through line 43, the first feature tree and the second feature tree are merged to form a third feature tree according to the same coordinate position. In addition, when storing the data of the third feature tree, the number of the first feature tree and the number of the second feature tree are recorded, and the merging function of the first feature tree and the second feature tree is calculated to form a general feature tree with direction.
[0067] Since there are a large number of elements on the layout that match the first feature tree, the second feature tree, and the third feature tree, the resistance data of these feature trees are the same, but the coordinates and directions of the center electrodes are different. In order to reduce the amount of data storage, it is not necessary to record the resistance value data of each element. Instead, it is necessary to record the data of a feature tree and record the center point coordinates and direction data of each element.
[0068] Finally, check whether all elements of the power supply network have been divided. If not, continue to divide the power supply network of the device layer into an octree structure. If the division is completed, end the division of this layer and proceed to the next layer. It is also necessary to record the depth of the octree structure data NETi_Tree_data of the completed layer mapped to the unit network and store the data in a multi-level storage system.
[0069] By dividing a single element into an octree structure using the above method, the complexity of building an octree structure can be reduced, and the relationship between tree nodes can be easily obtained, thereby forming a more efficient hierarchical data storage system. Similarly, the division of the ground network is the same as that of the power network. Compared with the power network, Figure 5 In the example shown, the ground network may have a larger polygonal irregular area. Therefore, it is necessary to first divide the larger area into multiple regular elements and form a small regular feature tree.
[0070] Then, step S14 is executed, and according to the result of the octree structure division of each power network, the data NETi_Tree_data of the octree structure of the power network is mapped and stored in the multi-level storage system, and the resistance value of each node in the octree structure needs to be calculated according to the conductivity of different material areas in the integrated circuit, and the resistance value of each node is stored in the multi-level storage system. The multi-level storage system is divided into a high-speed storage area and a low-speed storage area according to hardware. Preferably, this embodiment stores each unit divided by the octree structure, including the specific feature tree subnode, in the low-speed storage area by optimizing the memory and parallel division.
[0071] Since each unit divided by the octree structure retains a root node, each child node is arranged in order, which is convenient for subsequent addition, deletion, modification and query. In addition, the octree structure is essentially a hierarchical data structure, which constructs the entire tree structure by recursively dividing each spatial area into smaller sub-areas. In this embodiment, the characteristics of the data make it mainly store regular structural units, which are formed based on the minimum division matrix unit Matrix_Min_i calculated in step S12. Since the shape, size and division method of the regular unit have clear regularity, it can be directly mapped to each node of the octree structure, thereby forming a clear hierarchical relationship.
[0072] In contrast, an irregular unit is a unit whose topological structure of the division unit does not have a clear regularity. In a complex system-level circuit structure, irregular units are common. This embodiment transforms irregular units into regular units by constructing virtual regular octree node units and stores data. This division is relatively simple and easy to judge whether each element meets the division criteria of the regular unit. Therefore, in the process of constructing the octree structure, once an irregular unit is encountered, it is flexibly handled by introducing a virtual regular octree node unit or a feature tree node.
[0073] In three-dimensional space, each chip may have a stacked structure, which can be reflected in the node information of the octree structure by recording the mapping layer and the node depth of the data. The recording of this hierarchical information not only helps to clearly display the organizational structure of the data, but also facilitates the subsequent data addition, deletion, modification and query operations. Through the hierarchical and depth information in the node information, the target node that needs to be searched or processed can be quickly located and the corresponding operation can be performed.
[0074] Compared with the traditional octree structure division method, the octree structure division method adopted in this embodiment not only adds the processing method of virtual regular octree node units and feature tree nodes, but also each octree has multiple subtrees. The multiple subtrees are spliced through a deep search algorithm and stored by mapping IDs, which is conducive to the re-construction of the octree structure when the subtree structure changes due to local modification of the layout.
[0075] Moreover, since each subtree root node has the ID of the element, including the ID of the power network where the element is located, the resistance value ID, and the divided unit serial number ID, it is more preferred that when calculating irregular units, each formula used has its own ID, and the formula ID forms function pointer information, through which various data can be quickly searched and modified.
[0076] In step S14, after the octree structure of the power supply network is constructed, the resistance of each node on the octree structure is calculated according to the material conductivity, and the resistance value of each node is stored in a Figure 3 In the multi-level storage system shown in FIG. 1 , preferably, each node of the octree structure can be merged using the component merging method, for example, resistors connected in series are merged into one resistor, and in this way, a resistor model of an integrated circuit can be constructed, and each resistor node has a subtree ID of the node, which is convenient for calculating the resistance value of a certain unit of a certain PIN pin.
[0077] Finally, step S15 is executed to display the octree structure data NETi_Tree_data of a power network and display the octree structure of the power network and the resistance value of each node when the octree structure data NETi_Tree_data of a power network to be displayed is received.
[0078] The present invention has advantages in the following aspects: First, when constructing the octree structure, full use is made of the processing method of the virtual rule octree node unit and the feature tree node, making the construction of the octree structure simpler and more flexible. In addition, by utilizing the hierarchical structure and parallel processing, combined with the stacked hierarchical division of the integrated circuit layout and the hierarchical storage structure of the octree structure, a parallel execution query algorithm is designed, which significantly reduces the computational complexity and time consumption, especially when processing large-scale data sets, and can significantly improve the processing speed and efficiency. In addition, the present invention can avoid repeated division of the octree structure for elements with the same characteristics. For elements with the same feature nodes, the merge process is directly performed by translating the center node of the tree, avoiding repeated subtree division and reducing the amount of calculation.
[0079] Second, the present invention adopts an efficient storage structure, and minimizes the space occupied by the octree structure during the storage process by designing a more compact and efficient storage structure. In addition, the data of the octree structure is compressed using an advanced compression algorithm to further reduce the storage space requirement. In addition, the data storage layout is optimized and the utilization of storage space is improved through the design of hierarchical storage strategy and disk mapping nodes.
[0080] Second, the present invention can improve the accuracy of boundary processing of the power supply network. Since the present invention adopts a more sophisticated boundary division strategy, it ensures that the octree structure can maintain a high accuracy when processing shapes with complex boundaries. In addition, the present invention uses virtual regular octree node units to achieve the fitting of regular and irregular nodes, so as to more accurately represent and approximate complex boundaries and improve the processing effect of boundaries. In addition, the ID of the calculation formula for converting irregular units into regular units in the node data of the octree structure is used, and parallel processing of converting parallel irregular units into regular units is realized by means of function pointers, supporting flexible accuracy adjustment to meet the accuracy requirements of different application scenarios.
[0081] Third, the present invention has high adaptability. When calculating the hexahedral feature threshold of each power network, it allows the hexahedral equal division coefficient to be enlarged by integer multiples according to the needs of the actual application scenario, thereby designing a more flexible octree structure division method, which can realize automatic division of the octree structure according to the actual situation of the input file. In addition, the present invention can also establish a parameter adjustment mechanism, allowing the user to dynamically adjust the division parameters of the octree structure according to data characteristics and processing requirements, so as to improve the adaptability and flexibility of the technology. Due to the existence of the feature tree, the division accuracy is improved. This method can effectively improve the division performance.
[0082] Computer device embodiment:
[0083] The computer device of this embodiment can be various types of electronic devices, such as smart terminal devices, which have a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an information processing program for implementing the above-mentioned information processing method. When the processor executes the computer program, it implements the various steps of the above-mentioned DC resistance model extraction and storage method.
[0084] For example, a computer program may be divided into one or more modules, one or more modules are stored in a memory and executed by a processor to complete each module of the present invention. One or more modules may be a series of computer program instruction segments capable of completing a specific function, and the instruction segments are used to describe the execution process of the computer program in a terminal device.
[0085] It should be noted that the terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the schematic diagram of the present invention is only an example of a terminal device and does not constitute a limitation on the terminal device. The terminal device may include more or fewer components than shown in the figure, or may combine certain components, or different components. For example, the terminal device may also include an input / output device, a network access device, a bus, etc.
[0086] The processor referred to in the present invention may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.
[0087] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.
[0088] Storage medium embodiment:
[0089] If the computer program stored in the computer device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, each step of the above-mentioned DC resistance model extraction and storage method can be implemented.
[0090] Among them, the computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form, etc. Computer readable media may include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer readable media do not include electric carrier signals and telecommunication signals.
[0091] Finally, it should be emphasized that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. Method for extracting and storing DC resistance model, including: Acquire layout information of the integrated circuit to be analyzed, and divide the elements of the power network of the integrated circuit into an octree structure according to the layout information; Features: Calculate the minimum partition matrix unit of each power network, and divide each sub-node of each power network into an octree structure, and map and store the data of each power network into a multi-level storage system according to the division result of the octree structure; Calculating the resistance value of each node in the octree structure according to the conductivity of different material regions in the integrated circuit, and storing the resistance value of each node; After receiving the display instruction, display the octree structure of the integrated circuit and the resistance value of each node of the octree structure; Partitioning each sub-node of each power network into an octree structure includes: When dividing the octree structure for each of the child nodes, determine whether the unit of the current first-level octree structure is divided into the minimum division matrix unit, and if not, continue to divide the unit of the current first-level octree structure into the octree structure until the unit of the current first-level octree structure is divided into the minimum division matrix unit; Furthermore, for the irregular units on the octree structure, a virtual regular octree node unit is constructed by using the regular unit and a preset functional relationship, and the resistance value of the virtual regular octree node unit is calculated.
2. The method for extracting and storing a DC resistance model according to claim 1, characterized in that: The regular unit is a regular cube unit or a cuboid with square upper and lower surfaces, and the irregular unit is a unit other than a regular cube unit or a cuboid with square upper and lower surfaces.
3. The method for extracting and storing a DC resistance model according to claim 1 or 2, characterized in that: Dividing each sub-node of each power network into an octree structure further includes: After performing octree structure division on a target child node, the features of the target child node are recorded and a feature tree is formed. When performing octree structure division on subsequent child nodes, if the subsequent child nodes have the same features as the target child node, the octree structure of the target child node is directly used as the octree structure of the subsequent child node, and the coordinate offset between the subsequent child node and the target child node is recorded.
4. The method for extracting and storing a DC resistance model according to claim 3, characterized in that: Determining that the subsequent subnode has the same feature as the target subnode includes: the subsequent subnode has the same via radius and / or uses the same material as the target subnode.
5. The method for extracting and storing a DC resistance model according to claim 3, characterized in that: Dividing each subnode of each power supply network into an octree structure also includes: if the power supply network is a device layer power supply network, obtaining a first feature tree of a first type of element and a second feature tree of a second type of element, and then merging the first feature tree and the second feature tree according to overlapping nodes of the first type of element and the second type of element to form a third feature tree.
6. The method for extracting and storing a DC resistance model according to claim 5, characterized in that: When storing the data of the third feature tree, the serial number of the first feature tree and the serial number of the second feature tree are recorded, and a merging function of the first feature tree and the second feature tree is calculated to form a general feature tree with a direction.
7. The method for extracting and storing a DC resistance model according to claim 1 or 2, characterized in that: When calculating the minimum partitioning matrix unit of each of the power supply networks, the length of the minimum partitioning matrix unit is determined by the minimum length of each element of the power supply network, the width of the minimum partitioning matrix unit is determined by the minimum width of each element of the power supply network, and the height of the minimum partitioning matrix unit is set to a preset value.
8. The method for extracting and storing a DC resistance model according to claim 1 or 2, characterized in that: Mapping and storing the data of each power supply network into a multi-level storage system includes: mapping each node of the octree structure of the power supply network into a one-dimensional array in a linear index manner, and determining the position of each node in the array according to the position of the parent node of each node and the order of the child nodes of the node in the parent node.
9. A computer device, characterized in that The method comprises a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, each step of the method for extracting and storing a DC resistance model as claimed in any one of claims 1 to 8 is implemented.
10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for extracting and storing a DC resistance model as claimed in any one of claims 1 to 8 are implemented.
Citation Information
Patent Citations
Model display method, equipment and device, electronic equipment and storage medium
CN114529648A
Method for assisting game display effect, mobile terminal and computer readable storage medium
CN110917610A
Information processing method and electronic equipment
CN113867668A
Data processing method and apparatus, electronic device, and storage medium
WO2021120190A1
