Methods, apparatus, electronic devices and products for building on-chip networks

Abstract

This application discloses an on-chip network construction method, apparatus, electronic device, and product, relating to the field of computer technology. The method includes obtaining cross-layer network nodes in each of multiple network regions uniformly divided by a bottom-layer cellular network, where the cross-layer network nodes are determined based on the number of bottom-layer networks in the bottom-layer cellular network; determining the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes in each network region; obtaining the node connection relationships between the cross-layer network nodes and the upper-layer network nodes; and constructing an on-chip network structure interconnecting the bottom-layer cellular network and the upper-layer network based on the node connection relationships. This solves the technical problem of rapidly increasing network latency in cellular on-chip networks as the network scale increases, achieving the technical effects of reducing on-chip network latency and increasing network throughput.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to methods, apparatus, electronic devices and products for constructing on-chip networks. Background Technology

[0002] With the rapid development of integrated circuit technology, on-chip networks have become an indispensable part of current integrated circuit design. Cellular on-chip networks, through their efficient communication architecture, enable rapid data exchange between each component within the chip. However, as the scale of cellular on-chip networks continues to increase, the network diameter also increases, leading to a rapid increase in the overall network latency. Summary of the Invention

[0003] This application provides a method, apparatus, electronic device, and product for constructing on-chip networks, in order to at least solve the problem in the related art that as the scale of cellular on-chip networks continues to increase, the network diameter also increases, resulting in a rapid increase in the latency of the entire network.

[0004] This application provides a method for constructing an on-chip network, including:

[0005] Obtain the cross-layer network nodes for each of the multiple network regions uniformly divided by the underlying cellular network. The cross-layer network nodes are determined based on the number of underlying networks in the underlying cellular network.

[0006] Based on the cross-layer network nodes of each network region, determine the upper-layer network nodes for constructing the upper-layer network;

[0007] Obtain the node connection relationship between cross-layer network nodes and upper-layer network nodes;

[0008] Based on node connectivity, an on-chip network structure is constructed to interconnect the underlying cellular network with the upper-layer network.

[0009] In one optional implementation, obtaining cross-layer network nodes for each of multiple network regions uniformly divided by the underlying cellular network includes:

[0010] The number of underlying cellular networks is obtained, which is determined by the layer number of the cellular network obtained with the center of the underlying cellular network as the origin.

[0011] If the number of underlying networks is greater than 1, then any set of edges of the underlying cellular network with a number of 1 underlying networks is used as the symmetric extraction to divide the underlying cellular network into multiple network regions.

[0012] The number of candidate bottom-level nodes in each network region is counted, and based on the number and location of the candidate bottom-level nodes, the cross-layer network nodes in each network region are determined.

[0013] In one optional implementation, the number of candidate bottom-level nodes in each network region is counted, and based on the number and location of the candidate bottom-level nodes, the cross-layer network nodes for each network region are determined, including:

[0014] The number of candidate bottom-level nodes in each network region is determined based on the node positions of each of the multiple cellular networks on the same axis of symmetry.

[0015] Based on the number and location of candidate bottom-level nodes, cross-layer network nodes are determined for each network region, with the number of cross-layer network nodes being less than the number of candidate bottom-level nodes.

[0016] In one optional implementation, determining the number of candidate bottom-level nodes in each network region based on the node positions of each of the multiple cellular networks located on the same axis of symmetry includes:

[0017] One of the two axes of symmetry within each network region is taken as the region's axis of symmetry.

[0018] The number of candidate bottom-level nodes in each network region is determined based on the node location of each cellular network in multiple cellular networks located on the regional axis of symmetry.

[0019] In one alternative implementation, determining the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region includes:

[0020] The number of cross-layer network nodes in each network region is compared with the cross-layer node threshold to obtain the cross-layer node comparison result;

[0021] If the cross-layer node comparison result indicates that the number of cross-layer network nodes is less than or equal to the cross-layer node threshold, then it is determined that all cross-layer network nodes in each network region require a corresponding upper-layer network node.

[0022] Based on the number of upper-layer network nodes required for cross-layer network nodes in each network region, the upper-layer network nodes for constructing the upper-layer network are determined.

[0023] In one alternative implementation, determining the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region further includes:

[0024] If the cross-layer node comparison result indicates that the number of cross-layer network nodes is greater than the cross-layer network node threshold, then the first target cross-layer network node and / or the second target cross-layer network node for each network region need a corresponding upper-layer network node.

[0025] Based on the number of the first target cross-layer network nodes and / or the second target cross-layer network nodes in each network region, determine the upper-layer network nodes for constructing the upper-layer network.

[0026] In one optional implementation, obtaining the node connection relationship between the cross-layer network node and the upper-layer network node includes:

[0027] Based on the number of upper-layer network nodes, determine the upper-layer connection relationships between each upper-layer network node and the cross-layer connection relationships between each upper-layer network node and cross-layer network nodes. The node connection relationships include cross-layer connection relationships and upper-layer connection relationships.

[0028] In one optional implementation, determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes includes:

[0029] If the number of nodes in the upper-level network is equal to the number of network regions, then the number of upper-level networks in the upper-level network is determined to be 1.

[0030] Interconnect adjacent and opposite nodes in an upper-layer network with a maximum upper-layer network size of 1 through an additional high-speed path.

[0031] In one optional implementation, determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes further includes:

[0032] If the number of nodes in the upper-level network is greater than the number of network regions, then the number of upper-level networks is determined based on the ratio of the number of nodes in the upper-level network to the number of network regions.

[0033] The ring number of each upper-layer network node is determined based on the number of upper-layer networks, and upper-layer network nodes with ring numbers greater than 1 are interconnected with other adjacent upper-layer network nodes through additional high-speed paths.

[0034] In one optional implementation, the cross-layer connection relationship between each upper-layer network node and cross-layer network nodes in the upper-layer network is determined based on the number of upper-layer network nodes, including:

[0035] If the number of upper-layer networks in the upper-layer network is 1, then each upper-layer network node in the upper-layer network is connected to all cross-layer network nodes in the corresponding network area through an additional high-speed path.

[0036] In one optional implementation, determining the cross-layer connection relationship between each upper-layer network node and cross-layer network nodes in the upper-layer network based on the number of upper-layer network nodes further includes:

[0037] If the number of upper-layer networks in the upper-layer network is greater than 1, then each upper-layer network node in the upper-layer network is connected to the first target number of cross-layer network nodes and / or the second target number of cross-layer network nodes in the corresponding network area through an additional high-speed path.

[0038] In one optional implementation, based on node connectivity, an on-chip network structure is constructed to interconnect the underlying cellular network and the upper-layer network, including:

[0039] Based on the upper-layer connection relationships between each upper-layer network node and the cross-layer connection relationships between each upper-layer network node and cross-layer network nodes, an on-chip network structure is constructed to interconnect the lower-layer cellular network with the upper-layer network.

[0040] This application also provides an on-chip network construction apparatus, comprising:

[0041] The node acquisition module is used to acquire cross-layer network nodes in each of the multiple network regions uniformly divided by the underlying cellular network. The cross-layer network nodes are determined based on the number of underlying networks in the underlying cellular network.

[0042] The node determination module is used to determine the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region.

[0043] The relationship acquisition module is used to obtain the node connection relationship between cross-layer network nodes and upper-layer network nodes;

[0044] The network construction module is used to build an on-chip network structure that interconnects the underlying cellular network with the upper-layer network based on node connection relationships.

[0045] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing the steps of any of the above-described on-chip network construction methods when executing the computer program.

[0046] This application also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the above-described on-chip network construction methods.

[0047] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described on-chip network construction methods.

[0048] By dividing the underlying cellular network into multiple network regions evenly and determining the upper-layer network nodes based on the cross-layer network nodes in these regions, and then constructing an on-chip network structure that interconnects the underlying cellular network and the upper-layer network based on the connection relationship between the upper-layer network nodes and the cross-layer network nodes, nodes that are far apart in the on-chip network can communicate quickly through additional high-speed paths, effectively reducing latency. At the same time, the upper-layer network can offload traffic, reduce the congestion of the underlying cellular network, increase network throughput, and improve the communication efficiency and stability of the network. Attached Figure Description

[0049] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating the on-chip network construction method provided in this application embodiment;

[0051] Figure 2 A schematic diagram of the underlying cellular network provided in an embodiment of this application;

[0052] Figure 3 This is a schematic diagram of network region division provided in an embodiment of this application;

[0053] Figure 4 This is a schematic diagram illustrating the determination of cross-layer network nodes provided in an embodiment of this application;

[0054] Figure 5 A schematic diagram of the connection of an upper-layer network provided in an embodiment of this application;

[0055] Figure 6 A schematic diagram of the structure of an on-chip network provided in an embodiment of this application;

[0056] Figure 7 This is another schematic diagram of the structure of the on-chip network provided in the embodiments of this application;

[0057] Figure 8 Another flowchart illustrating the on-chip network construction method provided in this application embodiment;

[0058] Figure 9 This is a schematic diagram of the node connection relationship of the on-chip network provided in an embodiment of this application;

[0059] Figure 10 This is another connection diagram of the upper-layer network provided in an embodiment of this application;

[0060] Figure 11Another schematic diagram of the on-chip network provided in the embodiments of this application;

[0061] Figure 12 A structural block diagram of the on-chip network construction device provided in the embodiments of this application;

[0062] Figure 13 A schematic diagram of the structure of an electronic device is provided for an embodiment of this application. Detailed Implementation

[0063] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0064] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0065] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] This embodiment provides a method for constructing an on-chip network. Figure 1 This is a flowchart of an on-chip network construction method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0067] Step 110: Obtain the cross-layer network nodes of each of the multiple network regions uniformly divided by the underlying cellular network.

[0068] The aforementioned cross-layer network nodes are determined based on the number of underlying cellular networks. The underlying cellular network refers to a network structure formed by multiple cellular network cells (regular hexagons) arranged edge-to-edge. Each cellular network cell contains multiple network nodes (candidate underlying nodes). These network nodes are interconnected within the underlying cellular network via bidirectional links, forming the underlying communication network. Cross-layer network nodes are selected from these network nodes to construct the upper-layer network. By evenly dividing the underlying cellular network into multiple network regions and selecting cross-layer network nodes in each region, a more balanced and efficient construction of the upper-layer network can be ensured.

[0069] In some optional implementations, when obtaining the cross-layer network nodes of each network region in multiple network regions uniformly divided by the underlying cellular network, the number of underlying networks in the underlying cellular network can be obtained first. The number of underlying networks is determined by the layer number of the cellular network obtained with the center of the underlying cellular network as the origin. If the number of underlying networks is greater than 1, the underlying cellular network is divided into multiple network regions by taking any set of edges of the underlying cellular network with a number of 1 as the symmetric extraction. The number of candidate underlying nodes in each network region is counted, and the cross-layer network nodes of each network region are determined based on the number of candidate underlying nodes and node positions to ensure efficient and stable connection between the upper-layer network nodes and the underlying cellular network.

[0070] Specifically, the number of underlying cellular networks is determined by taking the center of the underlying cellular network as the origin and considering the number of layers extending outward from the underlying cellular network. For example, if the underlying cellular network contains only the center layer, the number of underlying networks is 1; if the underlying cellular network contains the center layer and the first outer layer, the number of underlying networks is 2, and so on.

[0071] For example, if the underlying cellular network consists of a single cell, i.e., a regular hexagon, then the number of underlying cells (network size) of the underlying cellular network is 1, denoted as HM1. Adding a hexagonal ring (i.e., 6 hexagons) to the boundary (six sides) of HM1 results in a cellular network with 2 underlying cells, denoted as HM2. Similarly, adding a hexagonal ring to the boundary of HM(t-1) results in a cellular network of size t, denoted as HMt. Figure 2 This is a cellular topology network for a two-dimensional HM3 underlying cellular network.

[0072] If the number of underlying networks is greater than 1, then when dividing the underlying cellular network into multiple network regions, any pair of edges of the underlying cellular network with a minimum of 1 network is used as the axis of symmetry. This divides the underlying cellular network into 6 network regions in the directions of up / down, top-left-bottom-right, and bottom-left-top-right. Each network region contains a certain number of candidate underlying nodes, which are interconnected through bidirectional links (underlying bidirectional links) to form the underlying communication network. This symmetry-based division ensures that the size and shape of each network region are relatively uniform, which is beneficial for the subsequent construction of upper-layer networks and communication between nodes.

[0073] For example, please refer to Figure 3 If the number of underlying networks is 3, then first establish a cellular network coordinate system, since it is perpendicular to a certain coordinate axis ( , or The coordinate values ​​of the sawtooth chain (axis) are equal, such as Figure 3 A thickened sawtooth chain, the sawtooth chain being perpendicular to... The axis, therefore, the nodes on this sawtooth chain The coordinate values ​​are equal. Figure 3 middle dashed line Using this boundary, the coordinate values ​​of the sawtooth chain perpendicular to the positive direction of the coordinate axis are positive, and counting begins from a coordinate value of 1 along the positive direction of the coordinate axis. Otherwise, the coordinate values ​​are non-positive, and counting begins from a coordinate value of 0 along the negative direction of the coordinate axis, such as... Figure 3 The nodes on the thickened black serrated chain The coordinate is equal to 1. In this embodiment, a dashed line is used. , , The cellular network is divided into Region I, Region II, Region III, Region IV, Region V, and Region VI.

[0074] Since HM1 has only 6 candidate bottom-layer nodes and a network diameter of 3, there is no need to build an upper-layer network. In this embodiment of the invention, the size of the bottom-layer cellular network (bottom network) is greater than 1, i.e. .

[0075] The structure based on the underlying cellular network is described by the following parameters: , , , , .

[0076] : Indicates the network type, with values ​​of 0 or 1. Determined by the size of the underlying cellular network (number of underlying networks). The network is divided into two categories: when When, it represents a class of networks, let ;when At that time, it represents another type of network, allowing .

[0077] : This represents the size of the underlying cellular network, derived from the definition of a cellular network. When When, use This represents the number of cross-layer network nodes connected to each upper-layer network node in the upper-layer network.

[0078] : Indicates the number of nodes in the upper-layer network.

[0079] This indicates the number of upper-layer network nodes that are connected to a specific network region in the lower-layer cellular network via additional high-speed channels. Therefore, there are a total of [number missing] nodes in the upper-layer network. Each upper-level network node is connected to four cross-level network nodes.

[0080] This indicates the number of upper-layer network nodes that are connected to a specific network region in the lower-layer cellular network via additional high-speed channels. Therefore, there are a total of [number missing] nodes in the upper-layer network. Each upper-level network node is connected to three cross-level network nodes.

[0081] for The network, composed of parameters , , To describe networks, i.e., hierarchical Network, indicating when the underlying cellular network size is At that time, each node in the upper-layer network is connected to the node in the lower-layer network. Several cross-layer network nodes are connected via additional high-speed channels, and the upper-layer network contains... Each upper-layer network node.

[0082] for The network, composed of parameters , , To describe networks, i.e., hierarchical Network, indicating when the underlying cellular network size is At that time, each network area of ​​the underlying cellular network has There are 10 upper-layer network nodes, each of which is connected to 3 cross-layer network nodes in the network area. Each network area of ​​the lower-layer cellular network has 10 upper-layer network nodes. There are four upper-level network nodes, each of which is connected to four cross-level network nodes in the network region.

[0083] In some optional implementations, when counting the number of candidate bottom-layer nodes in each network region and determining the cross-layer network nodes for each network region based on the number and location of the candidate bottom-layer nodes, the number of candidate bottom-layer nodes in each network region can be determined according to the node location of each cellular network in multiple cellular networks on the same axis of symmetry. The cross-layer network nodes for each network region are then determined based on the number and location of the candidate bottom-layer nodes, with the number of cross-layer network nodes being less than the number of candidate bottom-layer nodes. For example, if a network region contains 10 candidate bottom-layer nodes, 5 nodes can be selected as cross-layer network nodes based on factors such as node location and connection relationships. These cross-layer network nodes will be used to connect with upper-layer network nodes to build an efficient and stable communication network. In some embodiments, the cross-layer network nodes for each network region can be further optimized and adjusted according to actual needs to improve the performance and stability of the entire network.

[0084] In some optional implementations, when determining the number of candidate bottom-level nodes in each network region based on the node positions of each of the multiple cellular networks on the same axis of symmetry, one of the two axes of symmetry in each network region can be used as the regional axis of symmetry; the number of candidate bottom-level nodes in each network region is determined based on the node positions of each of the multiple cellular networks on the regional axis of symmetry.

[0085] Specifically, the number of nodes in the cellular network on the regional axis of symmetry can be averaged, or the number of nodes in the cellular network on the regional axis of symmetry can be determined based on statistical characteristics such as the median of the number of nodes in the cellular network on the regional axis of symmetry, so that there is at least one or one cross-layer network node in each cellular network, and the distance between two pairs of cross-layer network nodes in adjacent cellular networks is equal.

[0086] For example, nodes in the lower-level cellular network that are connected to the upper-level network via additional high-speed channels are defined as cross-layer network nodes. The cross-layer network nodes in each network region are:

[0087] Area I: The nodes are cross-layer network nodes;

[0088] Area II: The nodes are cross-layer network nodes;

[0089] Area III: The nodes are cross-layer network nodes;

[0090] Area IV: The nodes are cross-layer network nodes;

[0091] Area V: The nodes are cross-layer network nodes;

[0092] Area VI: The nodes are cross-layer network nodes.

[0093] Due to the above set and , Representing the set of positive integers, therefore, in the underlying cellular network Each region has There are [number] cross-layer nodes, and the entire network has a total of [number] nodes. A cross-layer network node. Figure 4 This is a schematic diagram of cross-layer network nodes in the underlying cellular network HM6. Only the cross-layer network nodes are shown in the diagram, using " "express," "" indicates a bidirectional link, and the intersection of a bidirectional link is a regular node or a candidate bottom-level node.

[0094] Step 120: Based on the cross-layer network nodes of each network region, determine the upper-layer network nodes for constructing the upper-layer network.

[0095] As shown above, by using cross-layer network nodes based on each network region, the upper-layer network nodes for building the upper-layer network are determined to form an on-chip network architecture with efficient communication and data processing capabilities.

[0096] In some optional implementations, when determining the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region, the number of cross-layer network nodes in each network region can be compared with a cross-layer node threshold to obtain a cross-layer node comparison result. If the cross-layer node comparison result indicates that the number of cross-layer network nodes is less than or equal to the cross-layer node threshold, then it is determined that all cross-layer network nodes in each network region require a corresponding upper-layer network node. Based on the number of upper-layer network nodes required by the cross-layer network nodes in each network region, the upper-layer network nodes for constructing the upper-layer network are determined.

[0097] Specifically, any upper-layer network node can only connect to cross-layer nodes within a single network area of ​​the underlying cellular network, and cannot connect to cross-layer network nodes in other network areas. Therefore, the number of upper-layer network nodes is determined by the number of cross-layer network nodes. After determining the correspondence between cross-layer network nodes and required upper-layer network nodes for each network area, the location information and connectivity of these cross-layer network nodes can be further analyzed to optimize the layout of upper-layer network nodes, ensuring that the communication links between upper-layer network nodes and cross-layer network nodes are as short as possible, thereby reducing communication latency and improving data transmission efficiency. When determining the layout of upper-layer network nodes, the distribution of cross-layer network nodes in the underlying cellular network and their relative positions can be considered to make the connections between upper-layer network nodes and cross-layer network nodes more efficient and reduce unnecessary communication overhead. For example, upper-layer network nodes can be placed in the center of the underlying cellular network or on the critical path to better serve cross-layer network nodes in various network areas. Simultaneously, appropriate resources and bandwidth can be allocated to upper-layer network nodes based on their functions and requirements to ensure the performance and stability of the entire on-chip network architecture.

[0098] For example, when In this context, the connection method between cross-layer network nodes and upper-layer network nodes in the underlying cellular network is as follows: Within a network area of ​​the underlying cellular network... Each cross-layer network node is connected to the same upper-layer network node.

[0099] From parameters , , To describe hierarchical cellular networks (on-chip networks), i.e., hierarchical Network, representing the size of the underlying cellular network. At that time, each node in the upper-layer network is connected to the node in the lower-layer cellular network. The cross-layer network nodes are connected via additional high-speed channels, and this Several cross-layer network nodes are in the same network region, and there are nodes in the upper layer network. Each upper-layer network node.

[0100] Due to the underlying cellular network There is A cross-layer network node, and when At that time, each node in the upper-layer network is connected to a node in the lower-layer cellular network. Since the nodes in the upper-layer network are connected across layers, the number of nodes in the upper-layer network is:

[0101]

[0102] Therefore, for the underlying cellular network ,when At that time, the number of nodes in the upper-layer network is always 6. As can be seen from the definition, at this time, Therefore, a network can also be directly represented as a hierarchical structure. .

[0103] Step 130: Obtain the node connection relationship between the cross-layer network node and the upper-layer network node.

[0104] As mentioned above, by obtaining the node connection relationships between cross-layer network nodes and upper-layer network nodes, the communication paths and data transmission methods between each node can be clarified. Since cross-layer network nodes act as a bridge between the underlying cellular network and the upper-layer network, their connection relationships with upper-layer network nodes directly affect the communication efficiency and data processing capabilities of the entire on-chip network architecture. Therefore, it is necessary to accurately obtain these node connection relationships for subsequent network optimization and resource allocation.

[0105] Specifically, node connectivity can include parameters such as the connection status, connection strength, and communication latency between cross-layer network nodes and upper-layer network nodes. These parameters can be obtained through actual network testing or simulation experiments. Furthermore, based on the connection strength between cross-layer network nodes and upper-layer network nodes, upper-layer network nodes can be placed closer to critical cross-layer network nodes to reduce communication latency and improve data transmission efficiency. Simultaneously, appropriate bandwidth and resources can be allocated to upper-layer network nodes according to their functions and requirements to ensure the performance and stability of the entire on-chip network architecture.

[0106] In some optional implementations, when obtaining the node connection relationship between cross-layer network nodes and upper-layer network nodes, the upper-layer connection relationship between each upper-layer network node in the upper-layer network and the cross-layer connection relationship between each upper-layer network node and cross-layer network nodes can be determined based on the number of upper-layer network nodes. The node connection relationship includes the cross-layer connection relationship and the upper-layer connection relationship.

[0107] In some optional implementations, when determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes, if the number of upper-layer network nodes is equal to the number of network regions, then the upper-layer network number of the upper-layer network is determined to be 1; adjacent and opposite upper-layer network nodes in the upper-layer network with an upper-layer network number of 1 are interconnected through additional high-speed paths to significantly reduce the data transmission latency between these nodes, while improving the bandwidth and stability of data transmission.

[0108] In some alternative implementations, the cross-layer connection relationship between each upper-layer network node and cross-layer network nodes in the upper-layer network is determined based on the number of upper-layer network nodes. This includes: if the number of upper-layer networks in the upper-layer network is 1, then each upper-layer network node in the upper-layer network is connected to all cross-layer network nodes in the corresponding network area through an additional high-speed path. This ensures that data can flow quickly between upper-layer network nodes and cross-layer network nodes. Whether it is uplink or downlink transmission, the low latency and high bandwidth advantages brought by the high-speed path can be enjoyed. Even if a cross-layer network node fails, the data can still be transmitted smoothly through other paths, thereby ensuring the stable operation of the on-chip network system.

[0109] For example, when At that time, the upper-layer network had a total of 6 nodes, connected as follows: Figure 5 As shown, the upper network has a diameter of 2, and nodes are connected by additional high-speed paths. Figure 6 For hierarchical network, Figure 6 For hierarchical The internet. Figure 6 and Figure 7 middle" "This represents a cross-layer network node in the underlying cellular network." "Indicates a regular node / candidate underlying node in the underlying cellular network," "" indicates a node in the upper-layer network. "Indicates a bidirectional link in the underlying cellular network," "Indicates an additional high-speed lane."

[0110] Step 140: Based on the node connection relationship, construct an on-chip network structure that interconnects the underlying cellular network with the upper-layer network.

[0111] As shown above, by constructing an on-chip network structure that interconnects the underlying cellular network with the upper-layer network based on node connectivity, effective communication and data transmission between the underlying cellular network and the upper-layer network can be achieved. This interconnected on-chip network structure not only improves communication efficiency but also enhances data processing capabilities, making the entire on-chip network architecture more efficient and stable.

[0112] Specifically, based on the upper-layer connection relationships between each upper-layer network node in the upper-layer network, and the cross-layer connection relationships between each upper-layer network node and cross-layer network nodes, an on-chip network structure is constructed to interconnect the lower-layer cellular network with the upper-layer network.

[0113] In hierarchical In the architecture, the topology of the upper-layer network nodes adopts a hexagonal ring connection pattern. Each upper-layer network node establishes a full-duplex communication link with its clockwise adjacent node and the node along the axis of symmetry. In specific implementation, when the parameter t≥2, a three-level connection system is formed among the upper-layer network nodes: the main link uses wavelength division multiplexing technology to achieve a transmission bandwidth of 10Gbps, the auxiliary link provides a 5Gbps redundant channel through time division multiplexing technology, and the emergency link is configured with a dynamically tunable laser to achieve a burst transmission capability of 100Gbps established on demand.

[0114] In hot-swappable maintenance scenarios, when upper-layer network node hardware needs to be replaced, the faulty node migrates its connection to adjacent nodes via a blockchain consensus protocol. The migration process involves first topology verification by nodes in regions I and IV, followed by routing table synchronization by nodes in regions II and V, and finally, latency parameter confirmation by nodes in regions III and VI. This ensures that during the maintenance of a single upper-layer network node, end-to-end transmission latency fluctuations are controlled within 3ns, and the bit error rate remains below 10%. -12 .

[0115] In summary, the on-chip network construction method provided by this invention enables distant nodes in the on-chip network to communicate quickly through additional high-speed paths, effectively reducing latency. Simultaneously, the upper-layer network can offload traffic, reducing congestion in the underlying cellular network and increasing network throughput. Furthermore, as the scale of the underlying cellular network increases, it can more significantly reduce the overall network diameter and average network distance, thereby reducing network latency.

[0116] Figure 8 A flowchart illustrating another embodiment of the on-chip network construction method of the present invention is shown. Figure 8 As shown, the method includes the following steps:

[0117] Step 810: Obtain the cross-layer network nodes of each of the multiple network regions uniformly divided by the underlying cellular network.

[0118] Please see details Figure 1 Step 110 of the illustrated embodiment will not be described again here.

[0119] Step 820: Based on the cross-layer network nodes of each network region, determine the upper-layer network nodes for constructing the upper-layer network.

[0120] Specifically, step 820 above includes:

[0121] Step 8201: Compare the number of cross-layer network nodes in each network region with the cross-layer node threshold to obtain the cross-layer node comparison result;

[0122] Step 8202: If the cross-layer node comparison result indicates that the number of cross-layer network nodes is greater than the cross-layer network node threshold, then it is determined that the first target cross-layer network node and / or the second target cross-layer network node of each network region need a corresponding upper-layer network node.

[0123] Step 8203: Based on the number of the first target cross-layer network nodes and / or the second target cross-layer network nodes in each network region, determine the upper-layer network nodes for constructing the upper-layer network.

[0124] Specifically, any upper-layer network node can only connect to cross-layer network nodes within a single network area of ​​the lower-layer cellular network, and cannot connect to cross-layer network nodes in other network areas. Therefore, the number of upper-layer network nodes is determined by the total number of the first and / or second target cross-layer network nodes, identified after screening and grouping. After determining these key cross-layer network nodes, their distribution and connection requirements can be further analyzed to precisely plan the location and number of upper-layer network nodes, ensuring that they effectively serve the cross-layer network nodes in each network area and optimizing the communication efficiency and data processing capabilities of the entire on-chip network.

[0125] For example, when In this case, the connection method between cross-layer network nodes and upper-layer network nodes in the lower-layer cellular network is as follows: one upper-layer network node is connected to four or three cross-layer network nodes in a network area of ​​the lower-layer cellular network. That is, the number of cross-layer network nodes in a region of the lower-layer cellular network is the sum of multiples of four and multiples of three. For the lower layer... The further away from the center of the underlying cellular network, the more nodes there are on the hexagonal ring, and the longer the routing distance. In order to balance the load, from the center of the underlying cellular network outward, four cross-layer network nodes (if the multiple of 4 is not 0) are first connected to one upper-layer network node, and after arranging the multiples of 4, the multiples of 3 are arranged.

[0126] The number of nodes in the upper-layer network is:

[0127]

[0128]

[0129]

[0130]

[0131] in, This indicates the number of upper-layer network nodes connected to three cross-layer network nodes within a region of the underlying cellular network. This indicates the number of upper-layer network nodes connected to four cross-layer network nodes within a network region of the underlying cellular network. This indicates the number of nodes in the upper-level network.

[0132] Furthermore, when At that time, hierarchical cellular networks are determined by parameters , , To describe, that is, hierarchical Network, representing the size of the underlying cellular network. At that time, for each network area of ​​the underlying cellular network, the upper layer network has One upper-level network node is connected to four cross-level network nodes. One upper-level network node is connected to three cross-level network nodes. Starting from the first hexagonal ring closest to the bottom-level network, one upper-level network node is first connected to four cross-level nodes, and so on, until the top-level network node is connected to the bottom-level network node. After connecting all upper-layer network nodes Each layer of the network is connected to a third layer of the network, and then a third layer of the network is connected to a third layer of the network, until all the layers of the network in the region are connected.

[0133] Example: For hierarchical HM8 For a network, specifically a bottom-layer cellular network with 8 network elements, calculate the number of upper-layer networks:

[0134]

[0135]

[0136]

[0137]

[0138] Therefore, this underlying cellular network can be represented as a hierarchy. In a network, for each network area of ​​the underlying cellular network, there is one upper-layer network node connected to four cross-layer network nodes, and one upper-layer network node connected to three cross-layer network nodes. Figure 9 For the underlying cellular network The diagram shows the connection between cross-layer network nodes and upper-layer network nodes in Region I. The connection methods of the other five regions are the same as those in Region I.

[0139] Step 830: Obtain the node connection relationship between the cross-layer network node and the upper-layer network node.

[0140] In some optional implementations, determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes further includes: if the number of upper-layer network nodes is greater than the number of network regions, determining the number of upper-layer networks based on the ratio of the number of upper-layer network nodes to the number of network regions; determining the ring number of each upper-layer network node based on the number of upper-layer networks, and interconnecting upper-layer network nodes with ring numbers greater than 1 with other adjacent upper-layer network nodes through additional high-speed paths to form a multi-layer on-chip network structure. The design of the additional high-speed paths aims to improve the data transmission efficiency between upper-layer network nodes at different levels, ensuring rapid information flow and processing.

[0141] In some optional implementations, determining the cross-layer connection relationship between each upper-layer network node and cross-layer network nodes in the upper-layer network, based on the number of upper-layer network nodes, further includes: if the number of upper-layer networks in the upper-layer network is greater than 1, then determining that each upper-layer network node in the upper-layer network is connected to a first target number of cross-layer network nodes and / or a second target number of cross-layer network nodes within the corresponding network area through an additional high-speed path to ensure fast data transmission and processing. Simultaneously, the multi-layered on-chip network structure can better support complex network applications and services, improving the flexibility and scalability of the entire system.

[0142] Specifically, if the number of upper-layer networks is greater than one, then nodes at different levels are interconnected through additional high-speed paths to form a multi-layered network architecture. This multi-layered architecture not only improves data transmission efficiency but also enhances the network's fault tolerance and stability. When a node at a certain level fails, data can be transmitted through other nodes, thus avoiding the impact of a single point of failure on the entire network system.

[0143] Furthermore, when determining the cross-layer connection relationships between upper-layer network nodes and cross-layer network nodes, cross-layer network nodes with special functions or high data transmission requirements can be connected to multiple upper-layer network nodes to improve data transmission bandwidth and stability. For ordinary cross-layer network nodes, they can be connected to only one upper-layer network node to simplify the network structure and reduce costs.

[0144] When implementing the on-chip network construction method provided by this invention, optimizations can be made based on specific application scenarios and requirements. For example, in data center or cloud computing environments, the efficiency and processing power of data transmission can be improved by increasing the number and layers of upper-layer network nodes. In Internet of Things (IoT) or sensor networks, greater emphasis can be placed on network stability and fault tolerance, and network reliability and stability can be improved by adding redundant connections and backup nodes.

[0145] For example, for The underlying network, the ring closest to the center of the upper network, i.e., the upper network node with ring number 1, adopts... Figure 5 The upper-layer network is interconnected in a manner that allows for the interconnection of each ring with six nodes. Therefore, the upper-layer network can be represented as follows: network, Indicates the number of rings. Figure 10 For the underlying cellular network HM10 , HM11 , HM12 , HM13 At that time, the connection method of the upper-layer network.

[0146] Example: For hierarchical For a network, calculate the number of nodes in the upper-layer network:

[0147]

[0148]

[0149]

[0150]

[0151] Therefore, the underlying cellular network is hierarchical. In a cellular network, for each network region at the bottom layer, there are two upper-layer network nodes connected to three cross-layer network nodes, such as... Figure 11 As shown in the figure, only cross-layer network nodes are marked in the bottom-layer cellular network, using " "express," "This indicates a bidirectional link in the underlying cellular network, and the intersection of the bidirectional link is a regular node in the underlying cellular network." "Indicates an upper-layer network node, " "Indicates an additional high-speed access route."

[0152] Characteristics of parameter-based hierarchical cellular networks:

[0153] Network diameter: refers to the maximum value of the shortest distance between all pairs of nodes in a network. The network diameter has a significant impact on network latency.

[0154] Node connectivity: refers to the number of links connecting a node to other nodes, reflecting the connectivity and path diversity of the structure.

[0155] For hierarchical network:

[0156] The upper-layer network has 6 nodes and 9 additional high-speed paths, and the hierarchical network has a total of Additional high-speed pathways; hierarchical The diameter of the network is Hierarchical In the network, the connectivity of ordinary nodes at the bottom layer is 2 or 3, the connectivity of nodes across layers is 4, and the connectivity of nodes at the top layer is... .

[0157] For hierarchical network:

[0158] Upper-layer networks have 1 node; when At that time, the upper-layer network has 21 additional high-speed paths, and the hierarchical network has a total of An additional high-speed pathway; the upper-layer network has Additional high-speed pathways, the hierarchical network has a total of Additional high-speed pathways; hierarchical The diameter of the network is ;level In the network, the connectivity of ordinary nodes at the bottom layer is 2 and 3, the connectivity of nodes across layers is 4, and the connectivity of nodes at the top layer is 6, 7, 8, and 9.

[0159] Underlying cellular network The network diameter is: ,when hour,

[0160]

[0161] when hour:

[0162]

[0163] And because ,so:

[0164]

[0165] As can be seen, the diameter of the parameter-based hierarchical cellular network (on-chip network) of this invention is smaller than that of a two-dimensional cellular network. Furthermore, as the scale of the cellular network increases, the hierarchical cellular network topology can more significantly reduce the network diameter and decrease network latency. The average network distance (the average transmission distance between all source and destination nodes in the network) reflects the average latency characteristics of the structure. Due to its computational complexity, it is not calculated here. However, when the network diameter of one topology is smaller than that of another, the average network distance is also smaller. Therefore, the average network distance of the parameter-based hierarchical cellular network proposed in this invention is smaller than that of a two-dimensional cellular network, and as the scale of the cellular network increases, the hierarchical cellular network topology can more significantly reduce the average network latency.

[0166] Step 840: Based on the node connection relationship, construct an on-chip network structure that interconnects the underlying cellular network and the upper-layer network.

[0167] Please see details Figure 1 Step 140 of the illustrated embodiment will not be described again here.

[0168] In summary, the on-chip network construction method provided by this invention uses an additional high-speed channel to connect the underlying cellular network and the upper-layer network. The underlying network is interconnected in the form of a cellular topology, and the upper-layer network is based on the underlying HM (Hybrid Network Model). The network size is determined by two different connection methods, using additional high-speed channels for interconnection. Nodes in the underlying cellular network are categorized into three types: ordinary / candidate underlying nodes, cross-layer network nodes, and upper-layer network nodes. This structure enables distant nodes in the on-chip network to communicate quickly via additional high-speed channels, effectively reducing latency. Simultaneously, the upper-layer network can offload traffic, reducing congestion in the underlying cellular network and increasing network throughput. Furthermore, as the scale of the underlying cellular network increases, the parameter-based hierarchical network topology proposed in this invention can more significantly reduce the network diameter and average network distance, thereby decreasing network latency. Therefore, the parameter-based hierarchical network architecture proposed in this invention can effectively solve the problems of communication latency and throughput performance degradation caused by the expansion of two-dimensional cellular on-chip network scale.

[0169] Figure 12 A schematic diagram of an embodiment of an on-chip network construction device according to the present invention is shown. Figure 12 As shown, the device includes:

[0170] The node acquisition module 1210 is used to acquire cross-layer network nodes in each of multiple network regions uniformly divided by the underlying cellular network. The cross-layer network nodes are determined based on the number of underlying networks in the underlying cellular network.

[0171] The node determination module 1220 is used to determine the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region.

[0172] The relationship acquisition module 1230 is used to acquire the node connection relationship between cross-layer network nodes and upper-layer network nodes;

[0173] Network building module 1240 is used to build an on-chip network structure that interconnects the underlying cellular network with the upper-layer network based on node connection relationships.

[0174] In one optional implementation, the node acquisition module 1210 includes:

[0175] The underlying network number acquisition submodule is used to obtain the underlying network number of the underlying cellular network. The underlying network number is determined by the layer number of the cellular network obtained with the center of the underlying cellular network as the origin.

[0176] The network region partitioning submodule is used to divide the underlying cellular network into multiple network regions if the number of underlying networks is greater than 1, using any set of edges of the underlying cellular network with a number of 1 underlying networks as the symmetric extraction.

[0177] The cross-layer network node determination submodule is used to count the number of candidate bottom-layer nodes in each network region, and determine the cross-layer network nodes in each network region based on the number of candidate bottom-layer nodes and their locations.

[0178] In one alternative implementation, the cross-layer network node determination submodule includes:

[0179] The candidate bottom-level node number determination unit is used to determine the number of candidate bottom-level nodes in each network region based on the node position of each cellular network in multiple cellular networks on the same axis of symmetry.

[0180] The cross-layer network node determination unit is used to determine the cross-layer network nodes for each network region based on the number and location of candidate bottom-layer nodes, wherein the number of cross-layer network nodes is less than the number of candidate bottom-layer nodes.

[0181] In one optional implementation, the candidate bottom-level node number determination unit includes:

[0182] The region symmetry axis determines the sub-unit, which is used to take one of the two symmetry axes in each network region as the region symmetry axis;

[0183] The layer network node number determination sub-unit is used to determine the number of candidate bottom layer nodes in each network region based on the node position of each of the multiple cellular networks located on the regional axis of symmetry.

[0184] In one optional implementation, the node determination module 1220 includes:

[0185] A cross-layer node comparison submodule is obtained, which is used to compare the number of cross-layer network nodes in each network region with the cross-layer node threshold to obtain the cross-layer node comparison result;

[0186] The first upper-layer node number determination submodule is used to determine that if the cross-layer node comparison result indicates that the number of cross-layer network nodes is less than or equal to the cross-layer node threshold, then all cross-layer network nodes in each network region need a corresponding upper-layer network node.

[0187] The first upper-layer node determination submodule is used to determine the upper-layer network nodes for constructing the upper-layer network based on the number of upper-layer network nodes required for cross-layer network nodes in each network region.

[0188] In one optional implementation, the node determination module 1220 further includes:

[0189] The second upper-layer node number determination submodule is used to determine that if the cross-layer node comparison result indicates that the number of cross-layer network nodes is greater than the cross-layer network node threshold, then the first target cross-layer network node and / or the second target cross-layer network node of each network region need a corresponding upper-layer network node.

[0190] The second upper-layer node determination submodule is used to determine the upper-layer network nodes for constructing the upper-layer network based on the number of the first target cross-layer network nodes and / or the number of the second target cross-layer network nodes in each network region.

[0191] In one optional implementation, the relationship acquisition module 1230 is specifically used to determine, based on the number of upper-layer network nodes, the upper-layer connection relationship between each upper-layer network node and the cross-layer connection relationship between each upper-layer network node and the cross-layer network node, wherein the node connection relationship includes the cross-layer connection relationship and the upper-layer connection relationship.

[0192] In one optional implementation, the relationship acquisition module 1230 includes:

[0193] The upper-layer network number determination submodule is used to determine the upper-layer network number as 1 if the number of upper-layer network nodes is equal to the number of network regions.

[0194] The upper-layer network node connection submodule is used to interconnect two adjacent and opposite upper-layer network nodes in an upper-layer network with a maximum number of one upper-layer network nodes through an additional high-speed path.

[0195] In one optional implementation, the upper-layer network number determination submodule is further configured to determine the upper-layer network number based on the ratio of the upper-layer network number to the network region number if the number of upper-layer network nodes is greater than the number of network regions.

[0196] The upper-layer network node connection submodule is also used to determine the ring number of each upper-layer network node based on the number of upper-layer networks, and to interconnect upper-layer network nodes with ring numbers greater than 1 with other adjacent upper-layer network nodes through additional high-speed paths.

[0197] In an optional implementation, the relationship acquisition module 1230 further includes:

[0198] The cross-layer network node connection submodule is used to determine that each upper-layer network node in the upper-layer network is connected to all cross-layer network nodes in the corresponding network area through an additional high-speed path if the number of upper-layer networks in the upper-layer network is 1.

[0199] In one optional implementation, the cross-layer network node connection submodule is further configured to determine, if the number of upper-layer networks in the upper-layer network is greater than 1, that each upper-layer network node in the upper-layer network is connected to the first target number of cross-layer network nodes and / or the second target number of cross-layer network nodes in the corresponding network area through an additional high-speed path.

[0200] In one optional implementation, the network construction module 1240 is specifically used to construct an on-chip network structure that interconnects the underlying cellular network with the upper-layer network based on the upper-layer connection relationship between each upper-layer network node in the upper-layer network and the cross-layer connection relationship between each upper-layer network node and the cross-layer network node.

[0201] For a description of the features in the embodiment corresponding to the on-chip network construction device, please refer to the relevant description of the embodiment corresponding to the on-chip network construction method, which will not be repeated here.

[0202] Embodiments of this application also provide an electronic device, such as... Figure 13 As shown, it includes a memory 1310 and a processor 1320. The memory 1310 stores a computer program, and the processor 1320 is configured to run the computer program to perform the steps in any of the above embodiments of the on-chip network construction method.

[0203] The electronic device also includes a communication interface 1330 for communicating with other devices or communication networks.

[0204] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above embodiments of the on-chip network construction method when run.

[0205] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0206] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the on-chip network construction method described above.

[0207] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the on-chip network construction method.

[0208] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0209] The foregoing has provided a detailed description of an on-chip network construction method, apparatus, electronic device, and product provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A method for constructing an on-chip network, characterized in that, include: Obtain cross-layer network nodes for each of multiple network regions uniformly divided by the underlying cellular network, wherein the cross-layer network nodes are determined based on the number of underlying networks of the underlying cellular network; Based on the cross-layer network nodes of each network region, determine the upper-layer network nodes for constructing the upper-layer network; Obtain the node connection relationship between the cross-layer network node and the upper-layer network node; Based on the node connection relationship, an on-chip network structure is constructed to interconnect the underlying cellular network and the upper-layer network; The step of obtaining the node connection relationship between the cross-layer network node and the upper-layer network node includes: Based on the number of upper-layer network nodes, determine the upper-layer connection relationship between each upper-layer network node in the upper-layer network, and the cross-layer connection relationship between each upper-layer network node and the cross-layer network node. The node connection relationship includes the cross-layer connection relationship and the upper-layer connection relationship. Determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes further includes: If the number of upper-layer network nodes is greater than the number of network regions, then the number of upper-layer networks is determined based on the ratio of the number of upper-layer network nodes to the number of network regions. The ring number of each upper-layer network node is determined based on the number of upper-layer networks, and the upper-layer network nodes with ring numbers greater than 1 are interconnected with other adjacent upper-layer network nodes through additional high-speed paths.

2. The method according to claim 1, characterized in that, The step of obtaining cross-layer network nodes for each of multiple network regions uniformly divided by the underlying cellular network includes: The number of underlying networks of the underlying cellular network is obtained, and the number of underlying networks is determined by the layer number of the cellular network obtained with the center of the underlying cellular network as the origin. If the number of underlying networks is greater than 1, then the underlying cellular network is divided into multiple network regions by taking any set of edges of the underlying cellular network with a number of 1 as the symmetric extraction. The number of candidate bottom-level nodes in each network region is counted, and the cross-layer network nodes in each network region are determined based on the number and location of the candidate bottom-level nodes.

3. The method according to claim 2, characterized in that, The process of counting the number of candidate bottom-level nodes in each network region and determining the cross-layer network nodes in each network region based on the number and location of the candidate bottom-level nodes includes: The number of candidate bottom-level nodes in each network region is determined based on the node positions of each of the multiple cellular networks on the same axis of symmetry. Based on the number and location of the candidate bottom-level nodes, the cross-layer network nodes for each network region are determined, wherein the number of cross-layer network nodes is less than the number of candidate bottom-level nodes.

4. The method according to claim 3, characterized in that, The step of determining the number of candidate bottom-level nodes in each network region based on the node positions of each of the multiple cellular networks located on the same axis of symmetry includes: One of the two axes of symmetry within each network region is taken as the region's axis of symmetry. The number of candidate bottom-level nodes in each network region is determined based on the node position of each of the multiple cellular networks located on the axis of symmetry of the region.

5. The method according to claim 1, characterized in that, The determination of the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region includes: The number of cross-layer network nodes in each network region is compared with the cross-layer node threshold to obtain the cross-layer node comparison result; If the cross-layer node comparison result indicates that the number of cross-layer network nodes is less than or equal to the cross-layer node threshold, then it is determined that all cross-layer network nodes in each network region require a corresponding upper-layer network node. The upper-layer network nodes for constructing the upper-layer network are determined based on the number of upper-layer network nodes required for the cross-layer network nodes in each network region.

6. The method according to claim 5, characterized in that, The step of determining the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region further includes: If the cross-layer node comparison result indicates that the number of cross-layer network nodes is greater than the cross-layer network node threshold, then it is determined that the first target cross-layer network node and / or the second target cross-layer network node of each network region require a corresponding upper-layer network node. Based on the number of the first target cross-layer network nodes and / or the second target cross-layer network nodes in each network region, the upper-layer network nodes for constructing the upper-layer network are determined.

7. The method according to claim 1, characterized in that, Determining the upper-layer connection relationship between each upper-layer network node based on the number of upper-layer network nodes includes: If the number of nodes in the upper-layer network is equal to the number of network regions, then the number of upper-layer networks in the upper-layer network is determined to be 1. Two adjacent and opposite nodes in the upper-layer network with an upper-layer network number of 1 are interconnected through an additional high-speed path.

8. The method according to claim 7, characterized in that, Determining the cross-layer connection relationship between each upper-layer network node and the cross-layer network node based on the number of upper-layer network nodes includes: If the number of upper-layer networks in the upper-layer network is 1, then each upper-layer network node in the upper-layer network is connected to all cross-layer network nodes in the corresponding network area through an additional high-speed path.

9. The method according to claim 1, characterized in that, The step of determining the cross-layer connection relationship between each upper-layer network node and the cross-layer network node based on the number of upper-layer network nodes further includes: If the number of upper-layer networks in the upper-layer network is greater than 1, then each upper-layer network node in the upper-layer network is determined to be connected to the first target cross-layer network nodes and / or the second target cross-layer network nodes in the corresponding network area through an additional high-speed path.

10. The method according to claim 6, characterized in that, The on-chip network structure for interconnecting the underlying cellular network and the upper-layer network based on the node connection relationship includes: Based on the upper-layer connection relationships between each upper-layer network node in the upper-layer network, and the cross-layer connection relationships between each upper-layer network node and the cross-layer network node, an on-chip network structure interconnecting the lower-layer cellular network and the upper-layer network is constructed.

11. An on-chip network construction device, characterized in that, include: The node acquisition module is used to acquire cross-layer network nodes in each of multiple network regions uniformly divided by the underlying cellular network. The cross-layer network nodes are determined based on the number of underlying networks in the underlying cellular network. The node determination module is used to determine the upper-layer network nodes for constructing the upper-layer network based on the cross-layer network nodes of each network region. The relationship acquisition module is used to acquire the node connection relationship between the cross-layer network node and the upper-layer network node; The step of obtaining the node connection relationship between the cross-layer network node and the upper-layer network node includes: Based on the number of upper-layer network nodes, determine the upper-layer connection relationship between each upper-layer network node in the upper-layer network, and the cross-layer connection relationship between each upper-layer network node and the cross-layer network node. The node connection relationship includes the cross-layer connection relationship and the upper-layer connection relationship. Determining the upper-layer connection relationship between each upper-layer network node in the upper-layer network based on the number of upper-layer network nodes further includes: If the number of upper-layer network nodes is greater than the number of network regions, then the number of upper-layer networks is determined based on the ratio of the number of upper-layer network nodes to the number of network regions. The ring number of each upper-layer network node is determined based on the number of upper-layer networks, and the upper-layer network nodes with ring numbers greater than 1 are interconnected with other adjacent upper-layer network nodes through additional high-speed paths. The network construction module is used to construct an on-chip network structure that interconnects the underlying cellular network and the upper-layer network based on the node connection relationship.

12. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the on-chip network construction method as described in any one of claims 1 to 10 when executing the computer program.

13. A computer program product, characterized in that, Includes computer instructions for causing a computer to perform the steps of the on-chip network construction method according to any one of claims 1 to 10.

Images