Interposer power network routing method and system

Abstract

The application discloses a method and system for wiring an interposer power supply network, wherein the method comprises the following steps: determining each grid of a connection layer; constructing a plurality of first connection lines in a first direction and a plurality of second connection lines in a second direction; merging pins adjacent in the first direction and the second direction into a group of connected pins in each grid; dividing each grid into a plurality of groups of grids; establishing a connection relationship between the first connection lines and each group of grids, so that the pins in each group of grids are connected to the first connection lines or all the first connection lines are allocated to the corresponding grid; establishing a connection relationship between the second connection lines and each group of grids, so that the pins in each group of grids are connected to the second connection lines or all the second connection lines are allocated to the corresponding grid; and wiring according to the pins or grids connected by each first connection line and each second connection line. The application can improve the performance of a corresponding chip integrated system.

Description

Technical Field

[0001] This application relates to the field of integrated circuit technology, specifically to an interposer power network routing method and system. Background Technology

[0002] With the continuous development of semiconductor technology, the complexity and functionality of chips are increasing exponentially. Constrained by factors such as the "area wall," the traditional monolithic integrated circuit (SoC) architecture is gradually failing to meet the demands of artificial intelligence applications and high-performance computing for high bandwidth, low latency, and low power consumption. Against this backdrop, chiplet technology has emerged as a breakthrough in chip technology development. This technology involves breaking down a complex monolithic chip into multiple modular sub-chips (chiplets), which are then interconnected through an interposer. This technology not only reduces design and manufacturing complexity but also significantly improves chip yield and reduces manufacturing costs.

[0003] Building upon this foundation, the application of 2.5D and 3D integration technologies has gradually become a focus in the industry, especially driven by data center and artificial intelligence applications. Chip-in-chip technology has matured and is being widely adopted. Compared to traditional SoC architectures, chip-in-chip architecture offers significant advantages. First, modular design can better address the increasingly complex functional requirements of chips. Second, the discrete design of chips can effectively improve manufacturing yield, as each chip can be independently designed, manufactured, and tested, avoiding the low yield issues that occur in large-size single-chip production. However, despite these advantages, challenges remain in practical applications. In 2.5D packaging, the PDN (Power Distribution Network) on the interposer is the lifeline of the entire system's power integrity. Its importance has risen from a traditional PCB-level consideration to one of the core design challenges determining the success or failure of packaging. A low-impedance, high-performance interposer PDN is the cornerstone for ensuring the efficient and stable collaborative operation of multiple advanced chips.

[0004] The inventors discovered that the design of the intermediate layer power network in 2.5D integrated systems faces severe challenges. For example, when multiple high-performance chips work simultaneously, the transient current demand is large. Traditional intermediate layer power distribution networks have high impedance and low noise tolerance, which can easily cause voltage drops and ground bounce noise, resulting in poor performance of the corresponding chip integrated system. Summary of the Invention

[0005] In view of this, this application provides an interposer power network routing method and system to solve the problem that traditional interposer power network routing schemes easily lead to poor performance of the corresponding chip integrated system.

[0006] This application provides a method for intermediary layer power network cabling, comprising the following steps:

[0007] Determine the individual grids of the connection layer;

[0008] Construct a plurality of first connecting lines along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other;

[0009] In each grid, adjacent pins in the first and second directions are merged into a group of connected pins;

[0010] Each of the aforementioned grids is divided into multiple groups of grids, wherein the groups of grids do not overlap with each other;

[0011] Establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are connected to the first connection line or all the first connection lines are assigned to the corresponding grids;

[0012] Establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids;

[0013] Routing is performed based on the pins or grids connected to each of the first and second connecting lines respectively.

[0014] Optionally, after establishing the connection relationship between the second connection line and each group of meshes, so that the pins in each group of meshes are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding meshes, the intermediate layer power network routing method further includes: identifying unassigned first connection lines and assigning each unassigned first connection line to the group of meshes farthest from the corresponding first connection line; identifying unassigned second connection lines and assigning each unassigned second connection line to the group of meshes farthest from the corresponding second connection line.

[0015] Optionally, the method for determining the distance between the first connecting line and a set of grids includes: obtaining the first connecting line that is closest to the first connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the first connecting line and the corresponding set of grids based on the distance between the obtained first connecting line and the first connecting line.

[0016] Optionally, the method for determining the distance between the second connecting line and a set of grids includes: obtaining the second connecting line that is closest to the second connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the second connecting line and the corresponding set of grids based on the distance between the obtained second connecting line and the second connecting line.

[0017] Optionally, after routing the pins or grids connected according to each of the first connection lines and each of the second connection lines, the intermediate layer power network routing method further includes: if there are pins without assigned connection lines, reducing the spacing between the connection lines and returning to the process of determining each grid of the connection layer until all pins are assigned corresponding connection lines.

[0018] Optionally, the routing based on the pins or grids connected by each of the first connecting lines and each of the second connecting lines includes: if the projection of a pin falls on the corresponding first connecting line and the second connecting line, then a through hole is provided at the center of the pin to connect the corresponding first connecting line and the second connecting line through the through hole; if the projection of a pin does not fall on the corresponding first connecting line and the second connecting line, then a perpendicular line is provided from the center of the pin to the corresponding first connecting line and the second connecting line to connect the corresponding first connecting line and the second connecting line through the perpendicular line.

[0019] Optionally, the intermediate layer power network routing method further includes: connecting the first connection line and the second connection line corresponding to the same set of meshes to each other.

[0020] Optionally, establishing the connection relationship between the first connecting line and each group of grids, so that the pins in each group of grids are all connected to the first connecting line or all the first connecting lines are assigned to the corresponding grids, includes: S151, counting the number of unconnected pins in each group of grids relative to the nearest unassigned first connecting line, obtaining the number of unconnected pins in each group of grids relative to the corresponding first connecting line, taking a group of grids with the number of unconnected pins and connecting it to the corresponding first connecting line, setting all pins in that group of grids to the connected state, and setting the first connecting line to the assigned state; S152, returning to step S151 until all pins in each group of grids are connected or all the first connecting lines are assigned.

[0021] Optionally, establishing the connection relationship between the second connection line and each group of grids, so that the pins in each group of grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids, includes: S161, counting the number of unconnected pins in each group of grids relative to the nearest unassigned second connection line, obtaining the number of unconnected pins in each group of grids relative to the corresponding second connection line, taking a group of grids with the number of unconnected pins and connecting it to the corresponding second connection line, setting all pins in that group of grids to the connected state, and setting the second connection line to the assigned state; S162, returning to step S161 until all pins in each group of grids are connected or all the second connection lines are assigned.

[0022] Optionally, determining each grid of the connection layer includes: identifying each pin of the connection layer and determining each grid based on the minimum bounding loop passing through the center of each pin.

[0023] This application also provides an intermediate layer power network cabling system, including:

[0024] The determination module is used to determine the individual grids of the connection layer;

[0025] A construction module is used to construct a plurality of first connecting lines along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other;

[0026] The merging module is used to merge adjacent pins in the first and second directions into a set of connected pins in each grid.

[0027] A partitioning module is used to divide each of the grids into multiple groups of grids, wherein the groups of grids do not overlap with each other;

[0028] The first establishment module is used to establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are connected to the first connection line or all the first connection lines are assigned to the corresponding grids.

[0029] The second establishment module is used to establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids.

[0030] The routing module is used to perform routing based on the pins or grids connected to each of the first connecting lines and each of the second connecting lines.

[0031] This application also provides an intermediate layer power network cabling device, including: a memory and a processor, wherein the memory stores an intermediate layer power network cabling program, and when the intermediate layer power network cabling program is executed by the processor, it implements the steps of any of the above intermediate layer power network cabling methods.

[0032] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described intermediate layer power network routing methods.

[0033] The above-described intermediate layer power network routing method and system, by determining each grid of the connection layer, constructs multiple first connection lines along a first direction and multiple second connection lines along a second direction. In each grid, adjacent pins in the first and second directions are merged into a group of connected pins to increase the connectivity of pins within the same connection layer and reduce the corresponding voltage drop. Each grid is divided into multiple groups of grids, and connection relationships are established between the first connection lines and each group of grids, so that pins in each group of grids are all connected to the first connection lines or all the first connection lines are assigned to the corresponding grids. Connection relationships are also established between the second connection lines and each group of grids, so that pins in each group of grids are all connected to the second connection lines or all the second connection lines are assigned to the corresponding grids. All interconnects are assigned to their corresponding grids to achieve optimal pin connectivity within each grid group with the fewest possible interconnects, and to prevent short circuits between different grid groups. Routing is then performed based on the pins or grids connected by each first and second interconnect, achieving effective interconnection between grid groups. This provides a low-impedance gridded power network for top and bottom pins through the first interposer layer corresponding to the first interconnect and the second interposer layer corresponding to the second interconnect, providing low-resistance, low-inductance current paths for multiple top and bottom corresponding chips. This suppresses IR voltage drop and power supply noise, ensures system power supply stability, significantly improves power integrity, and thus enhances the performance of the corresponding chip integrated system. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0035] Figure 1 This is a schematic flowchart of an intermediary layer power network routing method according to an embodiment of this application;

[0036] Figure 2 This is a schematic diagram showing the layers of an embodiment of this application;

[0037] Figure 3 This is a schematic diagram of a mesh according to an embodiment of this application;

[0038] Figure 4 This is a schematic diagram of the connection line allocation according to an embodiment of this application;

[0039] Figure 5 This is a schematic diagram of the connection line routing according to an embodiment of this application;

[0040] Figure 6 This is a schematic diagram of an intermediary layer power network cabling system according to an embodiment of this application;

[0041] Figure 7 This is a schematic diagram of the structure of an intermediate layer power network cabling device according to an embodiment of this application. Detailed Implementation

[0042] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. In the absence of conflict, the following embodiments and their technical features can be combined with each other.

[0043] The first aspect of this application provides an intermediate layer power network cabling method, which can be performed by an intermediate layer power network cabling device that needs to perform intermediate layer power network cabling.

[0044] Specifically, refer to Figure 1 As shown, the intermediate layer power network routing method includes the following steps S110 to S170.

[0045] S110, determine the individual meshes of the connection layer.

[0046] refer to Figure 2 As shown, the connection layer can include a top layer (also called the TOP layer) and a bottom layer (also called the BOTTOM layer). Each connection layer includes multiple pins. These pins, along with the intermediary layer between the top and bottom layers, can form a corresponding power distribution network (PND). For example... Figure 2 In the top and bottom layers shown, P represents the power supply pin, and G represents the ground pin. For example... Figure 2 As shown, the pins on the top layer can be used to connect to the chip or dies above the top layer by pointing upwards, and the pins on the bottom layer can be used to connect to the chip or dies below the bottom layer by pointing downwards.

[0047] Determining each grid in the connection layer includes: identifying each pin of the connection layer, and determining each grid based on the minimum bounding ring (e.g., minimum rectangle) passing through the center of each pin; that is, obtaining the center point coordinates of all pins of the corresponding connection layer, then taking the minimum bounding box corresponding to the center point coordinates, and determining the grid range based on the minimum bounding box. Setting a minimum bounding ring here can limit the routing of a grid to the pin range of that grid, thereby providing more routing resources for other grids.

[0048] S120, construct a plurality of first connecting lines (such as track) along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other. For example, the first direction can be a horizontal direction or a direction parallel to a set of boundaries of the connecting layer, and the second direction is a direction perpendicular to the first direction.

[0049] Optionally, the first and second connecting lines can be referenced. Figure 2 As shown, multiple first connection lines are parallel to each other and are disposed on a first interposer layer. The number of layers in the first interposer layer is determined by the number of first connection lines; it can be one layer or multiple layers. Multiple second connection lines are parallel to each other and are disposed on a second interposer layer. The number of layers in the second interposer layer can be equal to the number of layers in the first interposer layer. If the boundary of the connection layer is rectangular, the first connection line can be a line parallel to one set of boundaries in the corresponding boundary rectangle, and the second connection line can be a line parallel to another set of boundaries in the corresponding boundary rectangle. Optionally, the aforementioned first and second connection lines are currently virtual lines. Virtual first and second connection lines can be preset first. After the connection relationship between the first and second connection lines and the pins is determined, routing is performed according to the virtual first and second connection lines and their corresponding connection relationships. The first and second connection lines need to avoid the center point of the pins. That is, neither the first nor the second connection line passes through the center point of any pin, in order to reserve via positions for subsequent corresponding signal line pins, which connect to the corresponding signal line routing layer through the corresponding vias.

[0050] S130, in each grid, merges adjacent pins in the first and second directions into a group of connected pins.

[0051] Pins on the same connection layer can be arranged along a first direction or a second direction. Adjacent pins include two adjacent pins in the first direction or two adjacent pins in the second direction. Optionally, step S130 above can interconnect adjacent pins according to the connection through the first direction or the second direction, so that the interconnected pins are grouped together, thereby increasing the connectivity of pins within the same connection layer and reducing the corresponding voltage drop. If the first direction is horizontal and the second direction is vertical, adjacent pins can be interconnected through horizontal or vertical connections.

[0052] S140, each of the grids is divided into multiple groups of grids, wherein the groups of grids do not overlap.

[0053] Specifically, step S140 can divide meshes without overlapping regions (or intersections) into different groups of meshes, and group meshes with overlapping regions into one group, to obtain multiple groups of non-overlapping meshes. Here, the meshes in different groups do not overlap, making the routing resources independent between different groups of meshes, while meshes in the same group can share routing resources. For example, refer to... Figure 3 As shown, if a connecting line (such as...) Figure 3 Connector 1) is assigned to the first grid net1. Since the first grid net1 and the second grid net2 are in the same group, the second grid net2, which belongs to a different network within the same group, cannot use this connector. Because this connector passes through different groups (such as...) Figure 3 When groups 1 and 2 are different groups, they are not interconnected, so other different groups of meshes can continue to use this connection line.

[0054] S150, establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are connected to the first connection line or all the first connection lines are assigned to the corresponding grids.

[0055] The approach to allocating the first connection line in step S150 can include: prioritizing networks with more pins; if a certain group of meshes has the most unconnected pins near a candidate first connection line, the first connection line will be prioritized for that group of meshes, so as to use the fewest first connection lines to ensure the best pin connectivity within each group of meshes and to prevent short circuits between different groups of meshes.

[0056] In some examples, step S150 above, which establishes the connection relationship between the first connecting line and each group of the grids so that the pins in each group of the grids are correspondingly connected to the first connecting line or all the first connecting lines have been assigned to the corresponding grids, includes steps S151 and S152.

[0057] S151, count the number of unconnected pins in each group of grids relative to the nearest and unassigned first connection line, obtain the number of unconnected pins in each group of grids relative to the corresponding first connection line, take the group of grids with the number of unconnected pins and connect it to the corresponding first connection line, set all pins in the group of grids to the connected state, and set the first connection line to the assigned state.

[0058] S152, return to step S151 until all pins in each group of grids are connected or all first connection lines are assigned.

[0059] Optionally, in this example, when obtaining the first connection line corresponding to a certain pin, the first connection line between the center of the distance between the pin and the previous pin and the center between the pin and the next pin can be identified first, so as to prevent short circuits between interconnects of different pins.

[0060] S160, establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids.

[0061] The approach to allocating the second connection line in step S160 can include: prioritizing networks with more pins; if a certain group of meshes has the most unconnected pins near a candidate second connection line, the second connection line will be prioritized for that group of meshes, so as to use the fewest second connection lines to ensure the best pin connectivity within each group of meshes and to prevent short circuits between different groups of meshes.

[0062] In some examples, step S160 above, which establishes the connection relationship between the second connection line and each group of the grids so that the pins in each group of the grids are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding grids, includes steps S161 and S162.

[0063] S161, count the number of unconnected pins in each group of grids on the nearest and unassigned second connection line, obtain the number of unconnected pins in each group of grids relative to the corresponding second connection line, take the group of grids with the number of unconnected pins and connect it to the corresponding second connection line, set all pins in the group of grids to the connected state, and set the second connection line to the assigned state.

[0064] S162, return to step S161 until all pins in each grid group are connected or all second connection lines are assigned.

[0065] Optionally, in this example, when obtaining the second connection line corresponding to a certain pin, the second connection line between the center of the distance between the pin and the previous pin and the center between the pin and the next pin can be identified first, so as to prevent short circuits between interconnects of different pins.

[0066] S170, routing is performed according to the pins or grids connected by each of the first connecting lines and each of the second connecting lines respectively, so as to achieve effective interconnection between each group of grids.

[0067] Specifically, the pins or grids connected to each of the first and second connection lines have been assigned between the connection lines and pins. Step 170 above can output to a def file (standard data file) based on the pins or grids connected to each of the first and second connection lines, so as to perform actual routing according to this def file.

[0068] The aforementioned intermediate layer power network routing method, by determining each grid of the connection layer, constructs multiple first connection lines along a first direction and multiple second connection lines along a second direction. In each grid, adjacent pins in the first and second directions are merged into a group of connected pins to increase pin connectivity within the same connection layer and reduce corresponding voltage drops. Each grid is divided into multiple groups of grids, and connection relationships are established between the first connection lines and each group of grids, so that pins in each group of grids are correspondingly connected to the first connection lines, or the first connection lines are all assigned to their corresponding grids. Connection relationships are also established between the second connection lines and each group of grids, so that pins in each group of grids are correspondingly connected to the second connection lines, or all second connection lines are connected. All pins have been assigned to their corresponding grids to achieve optimal pin connectivity within each grid group with minimal wiring, ensuring no short circuits between different grid groups. Wiring is then performed based on the pins or grids connected by each first and second connection line, achieving effective interconnection between grid groups. This provides a low-impedance gridded power network for top and bottom pins via the first interposer layer corresponding to the first connection line and the second interposer layer corresponding to the second connection line. This provides low-resistance, low-inductance current paths for multiple top and bottom layer chips, suppressing IR voltage drop and power supply noise, ensuring system power supply stability, significantly improving power integrity, and ultimately enhancing the performance of the corresponding chip integrated system.

[0069] In some embodiments, after establishing the connection relationship between the second connection line and each group of the meshes, such that the pins in each group of the meshes are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding meshes, the intermediate layer power network routing method further includes steps S180 and S190.

[0070] S180: Identify unassigned first connection lines and assign each unassigned first connection line to the group of grids furthest from the corresponding first connection line, in order to allocate the remaining first connection lines, further strengthen the grid and reduce IR voltage drop.

[0071] In some examples, the distance between the first connecting line and a set of grids is determined by: obtaining the first connecting line that is closest to the first connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the first connecting line and the corresponding set of grids based on the distance between the obtained first connecting line and the first connecting line.

[0072] This example iterates through all unassigned first connections. For a given set of meshes, it retrieves all first connections assigned to that set and selects the closest first connection that is already assigned to that set. This distance is recorded as the distance between the set of meshes and the unassigned set. Then, it selects the set of meshes furthest from the unassigned first connection and assigns it to that unassigned first connection. For example, if adjacent first connections include track1, track2, track3, track4, and track5, with the following network distribution: track1 corresponds to the first mesh net1, track2 is unassigned, track3 corresponds to the first mesh net1, track4 corresponds to the second mesh net2, and track5 corresponds to the first mesh net1; then track2 is furthest from net2, so track2 is assigned to net2. This makes the mesh sizes more uniform across different sets of meshes, preventing excessive voltage drop caused by some meshes being too large.

[0073] S190, identify unassigned second connection lines, assign each unassigned second connection line to the group of grids furthest from the corresponding second connection line, in order to allocate the remaining second connection lines, further strengthen the grid and reduce IR voltage drop.

[0074] In some examples, the distance between the second connecting line and a set of grids is determined by: obtaining the second connecting line that is closest to the second connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the second connecting line and the corresponding set of grids based on the distance between the obtained second connecting line and the second connecting line.

[0075] This example iterates through all unassigned second connectors. For a given set of meshes, it retrieves all second connectors assigned to that set and selects the one closest to the unassigned second connector that is already assigned to that set. This distance is recorded as the distance between the set of meshes and the unassigned set. Then, it selects the set of meshes furthest from the unassigned second connector and assigns it to that unassigned second connector. This results in a more uniform mesh size across different sets, preventing excessive voltage drop caused by some meshes being too large.

[0076] In some examples, the spacing between any two adjacent first connection lines is equal, and the spacing between any two adjacent second connection lines is equal. After routing according to the pins or grids connected by each of the first connection lines and each of the second connection lines, the intermediate layer power network routing method further includes: if there are pins without assigned connection lines (including first connection lines and second connection lines), reducing the spacing between the connection lines and returning to the process of determining each grid of the connection layer until all pins are assigned corresponding connection lines. This example allows for dynamic adjustment of the spacing between connection lines. Optionally, the spacing between the first connection lines is equal to the spacing between the second connection lines.

[0077] In some embodiments, the routing based on the pins or grids connected by each of the first connection lines and each of the second connection lines includes steps S171 and S172.

[0078] S171, if the projection of a pin falls on the corresponding first connecting line and the second connecting line, then a through hole is provided at the center of the pin to connect the corresponding first connecting line and the second connecting line through the through hole.

[0079] Specifically, for the first connecting line, the first connecting line corresponding to the pin can be determined. If the first connecting line is inside the pin, that is, the projection of the pin falls on the first connecting line, a through hole can be directly drilled in the center of the pin, and the pin can be connected to the projection position of the pin on the first connecting line through the through hole. For the second connecting line, the second connecting line corresponding to the pin can be determined. If the second connecting line is inside the pin, that is, the projection of the pin falls on the second connecting line, a through hole can be directly drilled in the center of the pin, and the pin can be connected to the projection position of the pin on the second connecting line through the through hole.

[0080] S172, if the projection of a pin does not fall on the corresponding first connection line and the second connection line, then a perpendicular line is set from the center of the pin to the corresponding first connection line and the second connection line to connect the corresponding first connection line and the second connection line through the perpendicular line.

[0081] Specifically, for the first connection line, if the first connection line is outside the corresponding pin, i.e., the projection of the pin does not fall on the corresponding first connection line, then first, a line is drilled from the center of the pin in the layer (such as the top or bottom layer) (this line can be perpendicular to the corresponding first connection line), and connected to the corresponding position of the first connection line. Then, a through hole is drilled at this position to connect to the first connection line, so as to facilitate connecting the pin to the assigned first connection line. For the second connection line, if the second connection line is outside the corresponding pin, i.e., the projection of the pin does not fall on the corresponding second connection line, then first, a line is drilled from the center of the pin in the layer (such as the top or bottom layer) (this line can be perpendicular to the corresponding second connection line), and connected to the corresponding position of the second connection line. Then, a through hole is drilled at this position to connect to the second connection line, so as to facilitate connecting the pin to the assigned second connection line.

[0082] In some embodiments, the intermediate layer power network routing method further includes: connecting the first connection line and the second connection line corresponding to the same set of meshes to each other, thereby connecting the connection lines of each part of the same set of meshes.

[0083] Specifically, in this embodiment, all first connecting lines can be traversed, and all second connecting lines can be traversed according to the grid assigned to each first connecting line. If the grid assigned to a second connecting line is the same as that assigned to a first connecting line, a through hole can be made at the intersection of the two connecting lines to connect the two connecting lines through the through hole.

[0084] In some examples, reference Figure 4 As shown, Figure 4 The diagram shows the first intermediary layer, the second intermediary layer, the top layer, and the bottom layer.

[0085] In the first interposer layer, a series of parallel first connection lines are provided; in the second interposer layer, a series of parallel second connection lines are provided. At each intersection of the first and second connection lines, vias can be used to connect them vertically, forming a horizontally and vertically connected power grid covering the entire area, allowing current to flow freely. Each area has multiple power supply pins on the top and bottom layers. These pins can be connected to the first and second interposers using the above-described interposer layer power network routing method. If there is a corresponding connection line below or above a pin, a via is used to connect the pin to the corresponding connection line. If there is no connection line below or above a pin, a short line is first drawn on the layer where the pin is located, the nearest corresponding connection line is found, and then a via is used to connect the drawn line to the corresponding connection line. It should be noted that in this example, the power distribution network has at least two routing layers (such as one first interposer layer and one second interposer layer). The network density can also be increased to 4 or 6 layers as needed, and the number of interposer layers is always even.

[0086] Figure 4 The number of traces in the first and second interposer layers is determined by the size of the interposer layer. Each connecting trace has the same spacing. If the interposer layer is square, the number of traces on the first and second interposer layers can be equal; if the interposer layer is rectangular, the number of traces on the first and second interposer layers can be unequal. In some examples, the spacing between the traces in the first and second interposer layers can be set to be equal to achieve a more uniform network density. In other examples, the spacing between the traces in the first and second interposer layers can be different. Optionally, refer to... Figure 5 The diagram illustrates the allocation of the first connecting lines in the first direction. The horizontal lines represent multiple first connecting lines, namely connecting line 1, connecting line 2, and connecting line 3; the dots and numbers represent different grid pins, namely grid pin 1, grid pin 2, and grid pin 3. Among the grid pins in the first row, grid pin 1 has the most pins (2), while grid pin 2 has only 1 pin. According to the allocation principle, connecting line 1, which is closest to the first row pins, is assigned to grid pin 1; similarly, connecting line 2 is assigned to grid pin 2, and connecting line 3 is assigned to grid pin 3. The second connecting lines in the second direction can also be allocated in this manner.

[0087] In some examples, the spacing between connectors (including the spacing between first connectors and / or the spacing between second connectors) can be dynamically adjusted based on connectivity. If a pin is not connected, the spacing can be reduced and the number of traces increased to improve connectivity. Alternatively, the line width of the connectors can be dynamically adjusted based on IRdrop. If IRdrop is insufficient, the line width can be increased to improve IRdrop.

[0088] When the spacing between connectors is too large, it can easily lead to an insufficient number of connectors in that area, resulting in pins having no available connectors. To address this, after the routing algorithm characterized by the intermediate layer power network routing method is completed, connectivity analysis can be performed to identify all unconnected pins. If unconnected pins are found, the spacing between connectors is reduced, for example, from 2μm to 1.8μm. Based on the new, smaller spacing, the intermediate layer power network routing method is re-executed to generate a routing mesh. Connectivity analysis can then be performed again to identify any unconnected pins and adjust the spacing until all pins are successfully connected or the minimum spacing limit allowed by the process is reached.

[0089] The inventors discovered that even when all pins are assigned corresponding traces, thin traces (i.e., small linewidth) can lead to high resistance and insufficient voltage at the terminals under high current. Based on this research, in some examples, a detailed circuit model containing all trace resistances, inductances, and capacitances can be extracted from the existing power network. Power integrity simulation is then performed to calculate the voltage at each point in the network. Pins or areas with excessive IRdrop (e.g., voltage below 95% of the target value) are identified. If IRdrop fails, the trace width of traces on critical paths (such as high-current trunks) can be increased. After updating the design, the model is extracted again and simulated until all IRdrop parameters meet the requirements.

[0090] The above-described intermediate layer power network routing method, by defining each grid of the connection layer, constructs multiple first connection lines along a first direction and multiple second connection lines along a second direction. In each grid, adjacent pins in the first and second directions are merged into a group of connected pins to increase pin connectivity within the same connection layer and reduce corresponding voltage drops. Each grid is divided into multiple groups of grids, and connection relationships are established between the first connection lines and each group of grids, so that pins in each group of grids are correspondingly connected to the first connection lines, or the first connection lines are all assigned to their corresponding grids. Connection relationships are also established between the second connection lines and each group of grids, so that pins in each group of grids are correspondingly connected to the second connection lines, or all second connection lines are connected. All pins have been assigned to their corresponding grids to achieve optimal pin connectivity within each grid group with minimal wiring, ensuring no short circuits between different grid groups. Wiring is then performed based on the pins or grids connected by each first and second connection line, achieving effective interconnection between grid groups. This provides a low-impedance gridded power network for top and bottom pins via the first interposer layer corresponding to the first connection line and the second interposer layer corresponding to the second connection line. This provides low-resistance, low-inductance current paths for multiple top and bottom layer chips, suppressing IR voltage drop and power supply noise, ensuring system power supply stability, significantly improving power integrity, and ultimately enhancing the performance of the corresponding chip integrated system.

[0091] Furthermore, the aforementioned intermediate layer power network routing method, by setting a core design rule that power routing avoids the pin center, reserves a clean and shortest vertical channel for signal vias. This rule achieves decoupling and isolation between the power network and the signal network from a physical layout perspective, completely avoiding signal quality degradation caused by routing conflicts and ensuring the performance of high-speed interconnects. The proposed intelligent track allocation algorithm based on pin distribution, connectivity priority, and distance constraints can automatically and optimally allocate limited routing resources to different power networks, replacing tedious manual layout, significantly improving design efficiency, and obtaining a globally superior routing scheme, improving pin connectivity and network uniformity, and achieving high automation and optimized routing. Pin grouping and merging are performed automatically based on pin spatial position, supporting independent and shared routing resource management for different power domains, naturally adapting to complex scenarios with multiple voltage domains and multiple cores. It also supports dynamic adjustment of line spacing based on actual connectivity requirements and dynamic adjustment of line width based on IR voltage drop indicators, enabling the design to have self-optimization capabilities and cope with different performance and process constraints.

[0092] A second aspect of this application provides an intermediate layer power network cabling system, which can be installed in an intermediate layer power network cabling device that requires intermediate layer power network cabling. (Reference) Figure 6 As shown, the aforementioned intermediate layer power network cabling system includes:

[0093] Module 110 is used to determine the individual grids of the connection layer;

[0094] Construction module 120 is used to construct a plurality of first connecting lines along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other;

[0095] Merging module 130 is used to merge adjacent pins in the first and second directions into a set of connected pins in each grid.

[0096] The partitioning module 140 is used to divide each of the grids into multiple groups of grids, wherein the groups of grids do not overlap with each other;

[0097] The first establishment module 150 is used to establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are correspondingly connected to the first connection line or all the first connection lines have been assigned to the corresponding grids.

[0098] The second establishment module 160 is used to establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding grids.

[0099] The routing module 170 is used to perform routing based on the pins or grids connected to each of the first connecting lines and each of the second connecting lines.

[0100] Specific limitations regarding the intermediate layer power network cabling system can be found in the limitations of the intermediate layer power network cabling method described above, and will not be repeated here. Each module in the aforementioned intermediate layer power network cabling system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independent of the computing modules in the relevant computer equipment, or stored in software in the memory of the computer equipment, so that the computing modules can call and execute the operations corresponding to each of the above units.

[0101] This application also provides an intermediate layer power network cabling device, see reference. Figure 7 As shown, the intermediate layer power network routing device may include: a memory and a processor, wherein the memory stores an intermediate layer power network routing program, and when the intermediate layer power network routing program is executed by the processor, it implements the steps of the intermediate layer power network routing method as described in any of the above embodiments.

[0102] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the intermediate layer power network routing method as described in any of the above embodiments.

[0103] Although this application has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art based on a reading and understanding of this specification and the accompanying drawings. This application includes all such modifications and variations and is limited only by the scope of the appended claims. In particular, with respect to the various functions performed by the aforementioned components, the terminology used to describe such components is intended to correspond to any component (unless otherwise indicated) that performs the specified function of said component (e.g., is functionally equivalent to it), even if structurally not equivalent to the disclosed structure performing the functions in the exemplary implementations of this specification shown herein.

[0104] That is, the above description is only an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made using the content of this application’s specification and drawings, such as the combination of technical features between different embodiments, or direct or indirect application in other related technical fields, are similarly included within the patent protection scope of this application.

[0105] Furthermore, it should be understood that in the description of this application, the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Additionally, for structural elements with the same or similar characteristics, this application may use the same or different reference numerals for identification. Moreover, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0106] In this application, the term "exemplary" is used to mean "serving as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as more preferred or advantageous than other embodiments. This application has been provided above to enable any person skilled in the art to implement and use it. Various details have been set forth in the above description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be implemented without using these specific details. In other embodiments, well-known structures and processes will not be described in detail to avoid obscuring the description of this application with unnecessary detail. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

Claims

1. A method for routing intermediate layer power networks, characterized in that, The intermediate layer power network cabling method includes: Determine the individual grids of the connection layer; Construct a plurality of first connecting lines along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other; In each grid, adjacent pins in the first and second directions are merged into a group of connected pins; Each of the aforementioned grids is divided into multiple groups of grids, wherein the groups of grids do not overlap with each other; Establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are connected to the first connection line or all the first connection lines are assigned to the corresponding grids; Establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids; Routing is performed based on the pins or grids connected to each of the first connecting lines and each of the second connecting lines, respectively; After establishing the connection relationship between the second connection line and each group of meshes, so that the pins in each group of meshes are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding meshes, the intermediate layer power network routing method further includes: identifying unassigned first connection lines and assigning each unassigned first connection line to the group of meshes farthest from the corresponding first connection line; identifying unassigned second connection lines and assigning each unassigned second connection line to the group of meshes farthest from the corresponding second connection line; After routing the pins or grids connected according to each of the first connection lines and each of the second connection lines, the intermediate layer power network routing method further includes: if there are pins without assigned connection lines, reducing the spacing between the connection lines and returning to the process of determining each grid of the connection layer until all pins are assigned corresponding connection lines.

2. The intermediate layer power network cabling method according to claim 1, characterized in that, The method for determining the distance between the first connecting line and a set of grids includes: obtaining the first connecting line that is closest to the first connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the first connecting line and the corresponding set of grids based on the distance between the obtained first connecting line and the first connecting line. The method for determining the distance between the second connecting line and a set of grids includes: obtaining the second connecting line that is closest to the second connecting line and has been assigned to the corresponding network in the corresponding grid, and determining the distance between the second connecting line and the corresponding set of grids based on the distance between the obtained second connecting line and the second connecting line.

3. The intermediate layer power network cabling method according to claim 1, characterized in that, The routing based on the pins or grids connected by each of the first connecting lines and each of the second connecting lines includes: If the projection of a pin falls on the corresponding first connection line and the second connection line, a through hole is provided at the center of the pin to connect the corresponding first connection line and the second connection line through the through hole; If the projection of a pin does not fall on the corresponding first connection line and the second connection line, then a perpendicular line is set from the center of the pin to the corresponding first connection line and the second connection line to connect the corresponding first connection line and the second connection line.

4. The intermediate layer power network cabling method according to claim 1, characterized in that, The intermediate layer power network cabling method further includes: Connect the first and second connecting lines corresponding to the same set of grids.

5. The intermediate layer power network cabling method according to claim 1, characterized in that, The step of establishing the connection relationship between the first connecting line and each group of the grids, so that the pins in each group of the grids are correspondingly connected to the first connecting line or all the first connecting lines have been assigned to the corresponding grids, includes: S151, count the number of unconnected pins in each group of grids on the nearest and unassigned first connection line, obtain the number of unconnected pins in each group of grids relative to the corresponding first connection line, take the group of grids with the number of unconnected pins and connect it to the corresponding first connection line, set all pins in the group of grids to the connected state, and set the first connection line to the assigned state. S152, return to step S151 until all pins in each group of grids are connected or all first connection lines are assigned.

6. The intermediate layer power network cabling method according to claim 1, characterized in that, The step of establishing the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding grids, includes: S161, count the number of unconnected pins in each group of grids on the nearest and unassigned second connection line, obtain the number of unconnected pins in each group of grids relative to the corresponding second connection line, take the group of grids with the number of unconnected pins and connect it to the corresponding second connection line, set all pins in the group of grids to the connected state, and set the second connection line to the assigned state. S162, return to step S161 until all pins in each grid group are connected or all second connection lines are assigned.

7. The intermediate layer power network cabling method according to claim 1, characterized in that, The determination of each grid in the connection layer includes: Identify each pin of the connection layer and determine each grid based on the minimum bounding loop passing through the center of each pin.

8. A power network cabling system for intermediate layers, characterized in that, include: The determination module is used to determine the individual grids of the connection layer; A construction module is used to construct a plurality of first connecting lines along a first direction and a plurality of second connecting lines along a second direction, wherein the first direction and the second direction are perpendicular to each other; The merging module is used to merge adjacent pins in the first and second directions into a set of connected pins in each grid. A partitioning module is used to divide each of the grids into multiple groups of grids, wherein the groups of grids do not overlap with each other; The first establishment module is used to establish the connection relationship between the first connection line and each group of the grids, so that the pins in each group of the grids are connected to the first connection line or all the first connection lines are assigned to the corresponding grids. The second establishment module is used to establish the connection relationship between the second connection line and each group of the grids, so that the pins in each group of the grids are connected to the second connection line or all the second connection lines are assigned to the corresponding grids. A routing module is used to perform routing based on the pins or grids connected to each of the first connecting lines and each of the second connecting lines respectively; After establishing the connection relationship between the second connection line and each group of grids, so that the pins in each group of grids are correspondingly connected to the second connection line or all the second connection lines have been assigned to the corresponding grids, the method further includes: identifying unassigned first connection lines and assigning each unassigned first connection line to the group of grids farthest from the corresponding first connection line; identifying unassigned second connection lines and assigning each unassigned second connection line to the group of grids farthest from the corresponding second connection line. After routing the pins or grids connected by the first and second connecting lines respectively, the method further includes: if there are pins without assigned connecting lines, reducing the spacing between the connecting lines and returning to the process of determining each grid of the connection layer until all pins are assigned corresponding connecting lines.

Images