CN115865784A - Path generation method and device and electronic equipment - Google Patents

Path generation method and device and electronic equipment Download PDF

Info

Publication number
CN115865784A
CN115865784A CN202211471812.5A CN202211471812A CN115865784A CN 115865784 A CN115865784 A CN 115865784A CN 202211471812 A CN202211471812 A CN 202211471812A CN 115865784 A CN115865784 A CN 115865784A
Authority
CN
China
Prior art keywords
node
nodes
routing
target
routing node
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211471812.5A
Other languages
Chinese (zh)
Inventor
买廷瑞
刘建伟
梁喆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Aixin Technology Co ltd
Original Assignee
Beijing Aixin Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Aixin Technology Co ltd filed Critical Beijing Aixin Technology Co ltd
Priority to CN202211471812.5A priority Critical patent/CN115865784A/en
Publication of CN115865784A publication Critical patent/CN115865784A/en
Pending legal-status Critical Current

Links

Images

Abstract

The disclosure provides a path generation method, a path generation device and an electronic device, wherein the method comprises the following steps: generating a hybrid network structure according to the plurality of master device nodes, the plurality of slave device nodes and the plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and a target slave device node is determined, so that the situation of low communication transmission efficiency caused by a network structure can be avoided, the optimal access path between the target master device node and the target slave device node is effectively determined, and the communication transmission efficiency is improved.

Description

Path generation method and device and electronic equipment
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to a method and an apparatus for generating a path, and an electronic device.
Background
With the exponential growth of data traffic and the rapid development of intelligent devices, networks are more and more complex and diversified, and more factors, including stability, security, bandwidth, delay, load, etc., need to be considered. Therefore, as chip multiprocessor capabilities continue to increase, the communication efficiency of the network on chip is of critical importance to overall performance. In order to improve the communication efficiency, during the process that the master device accesses the slave device to transmit information, the optimal access path can be determined in the network structure for data transmission. Therefore, how to determine the access path in the network structure is very important.
Disclosure of Invention
The present disclosure is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, the present disclosure provides a path generation method, a path generation apparatus, and an electronic device, which may generate a hybrid network structure according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes, and may avoid a situation of low communication transmission efficiency caused by the network structure, and further, in a process that a target master device accesses a target slave device, at least one target routing node may be selected from the hybrid network structure, so that according to the at least one target routing node, an optimal access path between the target master device node and the target slave device node may be effectively determined, and communication transmission efficiency is improved.
According to a first aspect of the present disclosure, there is provided a path generation method, including: generating a hybrid network structure according to the plurality of master device nodes, the plurality of slave device nodes and the plurality of routing nodes; for a target master device node in the plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from the plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; and determining an access path between the target master node and the target slave node according to the at least one target routing node.
According to a second aspect of the embodiments of the present disclosure, there is provided a path generation apparatus including: the generation module is used for generating a hybrid network structure according to the plurality of master equipment nodes, the plurality of slave equipment nodes and the plurality of routing nodes; a first determining module, configured to determine, for a target master device node in the plurality of master device nodes, a target slave device node to be accessed by the target master device node from the plurality of slave device nodes; a second determining module, configured to determine at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; a third determining module, configured to determine, according to the at least one target routing node, an access path between the target master device node and the target slave device node.
According to a third aspect of the present disclosure, there is provided an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the path generation method provided by the embodiment of the first aspect of the disclosure.
According to a fourth aspect of the present disclosure, there is provided a computer-readable storage medium, wherein instructions, when executed by a processor of an electronic device, enable the electronic device to perform the path generation method set forth in the first aspect of the present disclosure.
According to a fifth aspect of the present disclosure, there is provided a computer program product, comprising a computer program, which, when executed by a processor of an electronic device, enables the electronic device to perform the path generation method set forth in the embodiments of the first aspect.
According to the technical scheme, a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and the target slave device node is determined, so that a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes, the situation that the communication transmission efficiency is low due to the network structure can be avoided, furthermore, in the process that the target master device accesses the target slave device, at least one target routing node can be selected from the hybrid network structure, and therefore according to the at least one target routing node, the optimal access path between the target master device node and the target slave device node can be effectively determined, and the communication transmission efficiency is improved.
Additional aspects and advantages of the disclosure will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The foregoing and/or additional aspects and advantages of the present disclosure will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic diagram of a crossbar network architecture shown in accordance with an exemplary embodiment;
FIG. 2 is a schematic diagram of a grid structure shown in accordance with an exemplary embodiment;
FIG. 3 is a flow diagram illustrating a method of path generation in accordance with an exemplary embodiment;
FIG. 4 is a flow diagram illustrating another path generation method in accordance with an exemplary embodiment;
FIG. 5 is a flow diagram illustrating another path generation method in accordance with an exemplary embodiment;
FIG. 6 is a flow diagram illustrating another path generation method in accordance with an exemplary embodiment;
FIG. 7 is a flow diagram illustrating a method of path generation in accordance with an exemplary embodiment;
FIG. 8 is a block diagram of a path generation apparatus shown in accordance with an exemplary embodiment;
FIG. 9 is a block diagram illustrating a path generation electronic device in accordance with an exemplary embodiment.
Detailed Description
Reference will now be made in detail to the embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and intended to be illustrative of the present disclosure, and should not be construed as limiting the present disclosure.
The network architecture of a network-on-chip (NoC) defines the physical layout of the distribution and connection of the various modules in the network on a chip. The selection of the network structure directly influences the network node degree, the network diameter and the network scale, thereby influencing the network delay, the throughput, the energy consumption, the area, the fault tolerance and the like, and finally generating important influence on the network performance parameters.
Currently, noC research typically involves several network structures: crossbar (Crossbar) and Mesh (Mesh) structures, wherein Crossbar is a fully connected structure as shown in fig. 1, and each node is directly connected to other nodes, which has the advantages of low delay, high bandwidth efficiency and simple structure. But the expandability is poor and the implementation cost is O (N) 2 ) Especially, when the number of nodes is large, the power consumption area is large, the congestion problem is serious, and the back end is difficult to realize. As shown in fig. 2, the Mesh is generally a two-dimensional topology structure, each routing node corresponds to a computing node, each routing node is connected with four nodes around it to form a two-dimensional network structure, and there are many paths from each node to another node. The advantages are good scalability, easy implementation of the back-end, implementation cost of O (sqrt (N)), but larger delay and lower bandwidth efficiency than Crossbar.
In order to solve the above problem, the present disclosure provides a path generation method and apparatus, and an electronic device.
A path generation method, a path generation device, an electronic apparatus, and a storage medium according to the present disclosure are described in detail below with reference to the accompanying drawings.
Fig. 3 is a flow diagram illustrating a path generation method according to an example embodiment. The path generation method is applicable to a path generation device. The path generating device may be, for example, a hardware device that can be connected to the chip through a bus or the like, or a controller in the hardware device, or control software in the hardware device, and may be set according to actual needs, which is not limited in this disclosure.
As shown in fig. 3, the path generation method includes the following steps:
step 301, a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes.
In the embodiment of the present disclosure, a plurality of master nodes, a plurality of slave nodes, and a plurality of routing nodes may be preset, where, in order to avoid communication collision, the number of slave nodes may be greater than or equal to the number of master nodes. For example, the number of master nodes may be 16, and the number of slave nodes may be greater than or equal to 16.
It should be noted that, in order to improve communication efficiency and avoid the problem of low communication efficiency caused by the network structure, a hybrid network structure may be generated according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes, where the number of routing nodes may be determined according to the number of master device nodes, the number of slave device nodes, and the hybrid network structure to be generated. For example, the number of master device nodes may be 16, the number of slave device nodes may be 16, and the number of routing nodes may be set to 32, where 16 routing nodes and 16 master device nodes generate a first network structure, 16 routing nodes and 16 slave device nodes generate a second network structure, and a hybrid network structure is generated according to the first network structure and the second network structure.
Step 302, aiming at a target master device node in the plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from the plurality of slave device nodes.
In order to implement data communication between the master device node and the slave device node, the master device node may access the slave device node, where the master device node that needs to access the slave device node is a target master device node, and the slave device node to be accessed is a target slave device node.
Step 303, determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node.
In order to improve communication efficiency, in the embodiment of the present disclosure, at least one target routing node through which the target master node accesses the target slave node may be determined from the hybrid network structure according to the target master node and the target slave node.
Step 304, determining an access path between the target master node and the target slave node according to at least one target routing node.
Furthermore, according to at least one target routing node through which the master device node accesses the target slave device node, an access path of the target master device node to the target slave device node can be determined. It should be noted that the access path of the target master node accessing the target slave node may be the shortest access path.
In summary, a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and a target slave device node is determined, so that a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes, the situation that the communication transmission efficiency is low due to the network structure can be avoided, furthermore, in the process that the target master device accesses the target slave device, at least one target routing node can be selected from the hybrid network structure, and therefore according to the at least one target routing node, the optimal access path between the target master device node and the target slave device node can be effectively determined, and the communication transmission efficiency is improved.
In order to improve communication efficiency and avoid a problem of low communication efficiency caused by a network structure, a hybrid network structure may be generated according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes, as shown in fig. 4, fig. 4 is a schematic flow diagram of another path generation method shown according to an exemplary embodiment, in an embodiment of the present disclosure, a first network structure may be generated by a master device node and a first routing node of the routing nodes, a second network structure may be generated by a slave device node and a second routing node of the routing nodes, and further, the first network and the second network may be connected to generate the hybrid network structure, and the embodiment shown in fig. 4 may include the following steps:
step 401, a first network structure is generated according to the plurality of master device nodes and a first routing node, which is matched with the plurality of master device nodes, in the plurality of routing nodes.
In order to improve the expandability of the network structure and reduce the implementation cost, a first network structure may be generated by a plurality of master device nodes and a first routing node in the plurality of routing nodes, and a second network structure may be generated by a plurality of slave device nodes and a second routing node in the plurality of routing nodes, where the first network structure and the second network structure may be the same network structure, and for example, the first network structure and the second network structure may be Mesh network structures.
It should be noted that, in the present disclosure, only the first network structure and the second network structure are taken as the same network structure for exemplary illustration, and the first network structure and the second network structure may also be different network structures, and the present disclosure is not limited specifically.
In order to avoid communication collision and improve communication efficiency, as a possible implementation manner of the embodiment of the present disclosure, a plurality of master device nodes may be sorted to obtain a sorting order of the plurality of master device nodes; determining a first routing node matched with each main equipment node in the plurality of main equipment nodes from the plurality of routing nodes according to the arrangement sequence of the plurality of main equipment nodes; grouping first routing nodes corresponding to the plurality of main equipment nodes according to the set number of the nodes to obtain a plurality of first groups; and creating a first network structure according to the first routing node in any one of the first groups and the main equipment node corresponding to the first routing node in any one of the first groups.
That is, device identification information (e.g., device numbers) of a plurality of master devices may be obtained first, then the plurality of master device nodes may be ranked according to the identification information of the plurality of master devices, a first routing node matching the ranked plurality of master device nodes is selected from the plurality of routing nodes, then the first routing nodes matching the plurality of master device nodes are grouped to obtain a plurality of first groups, the number of nodes in each first group may be the same, and a first network structure may be created for the first routing node and the master device node in each first group.
For example, with the number of the master devices being 16, the number of 16 master devices being m0 to m15, and the number of the first routing nodes matched with the sorted master device nodes being 16, the 16 routing nodes are divided into 4 groups, each group includes 4 routing nodes, and each group includes 4 routing nodes and 4 corresponding master device nodes, thereby forming a Mesh network.
Step 402, generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes.
In order to avoid communication collision and improve communication efficiency, as a possible implementation manner of the embodiment of the present disclosure, a second routing node matching each of a plurality of slave device nodes is determined from the plurality of routing nodes; grouping second routing nodes corresponding to the plurality of slave equipment nodes according to the number of the nodes to obtain a plurality of second groups; a second network fabric is created based on the second routing node in any second packet of the plurality of second packets and the slave node corresponding to the second routing node in any second packet.
That is, the plurality of slave device nodes may be randomly arranged out of order, and a second routing node matching the plurality of slave device nodes after random arrangement out of order may be selected from the plurality of routing nodes, and then, the second routing nodes matching the plurality of slave device nodes may be grouped to obtain a plurality of second packets, the number of nodes in each second packet may be the same, and a second network configuration may be created for the second routing node and the slave device node in each second packet.
For example, the number of slave device nodes is 16, the number of 16 slave device nodes is 0 to 15, the number of second routing nodes matched with the master device node after the unordered arrangement is 16, the 16 routing nodes are divided into 4 groups, each group includes 4 routing nodes, and each group includes 4 routing nodes and corresponding 4 slave device nodes, so as to form a Mesh network.
Step 403, connecting the first network structure and the second network structure to generate a hybrid network structure.
In order to reduce the complexity and communication delay of the network structure and improve the communication efficiency of the network structure, the first network structure and the second network structure may be connected to generate a hybrid network structure, for example, the first network structure and the second network structure are Mesh network structures, and the first network structure and the second network structure may be interconnected by using CrossBar.
As another example, a plurality of routing nodes are grouped to obtain a plurality of third packets, where each of the plurality of third packets includes a plurality of routing nodes, the plurality of routing nodes in each of the third packets are connected with each other by using a third network structure, and the plurality of third packets are connected with each other by using a fourth network structure; connecting a plurality of master device nodes and the plurality of slave device nodes through a third network fabric and a fourth network fabric to generate a hybrid network fabric.
For example, the number of routing nodes is S, the S routing nodes may be divided into N groups, where each group may include Qi routing nodes, where i is a positive integer not greater than N, the Qi routing nodes may be connected to each other by using a third network structure (e.g., mesh), and the N groups of routing nodes may be connected to each other by using a fourth network structure (e.g., crossbar). It should be noted that when S routing nodes are divided into N groups, the routing nodes may be divided equally or not, and the present disclosure is not limited specifically. When the S routing nodes are averagely divided into N groups, each group includes the same number of routing nodes, i.e., Q1= Q2= Q3=. = QN, and when the S routing nodes are unequally divided into N groups, the number of routing nodes included in each group may be different. Step 404, for a target master node in the plurality of master nodes, determining a target slave node to be accessed by the target master node from the plurality of slave nodes.
Step 405, determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node.
Step 406, determining an access path between the target master node and the target slave node according to the at least one target routing node.
It should be noted that the execution processes of steps 404 to 406 may be implemented by any one of the embodiments of the present disclosure, and the embodiments of the present disclosure do not limit this and are not described again.
In summary, a first network structure is generated according to the plurality of master device nodes and a first routing node, which is matched with the plurality of master device nodes, in the plurality of routing nodes; generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes; the first network structure and the second network structure are connected to generate a mixed network structure, so that the first network structure is generated by the plurality of main equipment nodes and the first routing nodes matched with the plurality of main equipment nodes, the second network structure is generated by the plurality of slave equipment nodes and the second routing nodes matched with the plurality of slave equipment nodes, the expandability of the network structure can be improved, the implementation cost is reduced, furthermore, the first network structure and the second network structure are connected to generate the mixed network structure, the complexity and the communication delay of the network structure can be reduced, and the communication efficiency of the network structure is improved.
In order to reduce the complexity and communication delay of the network structure and improve the communication efficiency of the network structure, as shown in fig. 5, fig. 5 is a flowchart illustrating another path generation method according to an exemplary embodiment, and in the embodiment of the present disclosure, a first network structure and a second network structure may be connected according to the number of first packets, the number of routing nodes in each first packet, the number of second packets, and the number of routing nodes in each second packet to generate a hybrid network structure, where the embodiment shown in fig. 5 may include the following steps:
step 501, generating a first network structure according to a plurality of master device nodes and a first routing node matched with the plurality of master device nodes in the plurality of routing nodes.
Step 502, generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes.
Step 503, obtaining a jth first routing node of an ith first packet in the multiple first packets corresponding to the multiple first network structures and an ith second routing node of a jth second packet in the multiple second packets corresponding to the multiple second network structures.
In order to avoid communication congestion, when a first network structure is connected with a second network structure, different routing nodes in each of a plurality of first packets corresponding to the first network structure may be connected with a plurality of second packets corresponding to the second network structure.
Step 504, the jth first routing node of the ith first packet is connected with the ith second routing node of the jth second packet to obtain the hybrid network structure.
And further, connecting the jth first routing node of the ith first packet with the ith second routing node of the jth second packet to obtain the hybrid network structure.
For example, the first network structure and the second network structure are Mesh network structures, the hybrid network structure to be generated is a Crossbar network structure, the number of first packets corresponding to the first network structure is 4, the number of first routing nodes in each first packet is 4, the number of second packets corresponding to the second network structure is 4, the number of second routing nodes in each second packet is 4, a first routing node in a first packet is connected to a first second routing node in a first second packet, and connecting the second first routing node in the first packet with the first second routing node in the second packet, connecting the third first routing node in the first packet with the first second routing node in the third second packet, and so on, and connecting the jth first routing node in the ith first packet with the ith second routing node in the jth second packet to obtain the hybrid network structure.
Step 505, for a target master node in the plurality of master nodes, determining a target slave node to be accessed by the target master node from the plurality of slave nodes.
Step 506, determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node.
Step 507, according to at least one target routing node, determining an access path between a target master node and a target slave node.
It should be noted that the execution processes of steps 501 to 502 and steps 505 to 507 may be implemented by any one of the embodiments of the present disclosure, and the embodiments of the present disclosure do not limit this and are not described again.
In summary, the jth first routing node of the ith first packet in the multiple first packets corresponding to the multiple first network structures and the ith second routing node of the jth second packet in the multiple second packets corresponding to the multiple second network structures are obtained; the jth first routing node of the ith first packet is connected with the ith second routing node of the jth second packet to obtain a hybrid network structure, so that different routing nodes in each first packet of a plurality of first packets corresponding to the first network structure are connected with different routing nodes in each second packet of a plurality of second packets corresponding to the second network structure, the hybrid network structure is generated, the complexity and communication delay of the network structure can be reduced, and the communication efficiency of the network structure is improved.
In order to accurately determine a path for a target master node to access a target slave node, as shown in fig. 6, fig. 6 is a flowchart of another path generation method shown in an exemplary embodiment, and in this embodiment of the disclosure, a routing node through which the target master node accesses the target slave node may be determined from a first packet and a second packet, and the passed routing node is taken as a target routing node, and the embodiment shown in fig. 6 may include the following steps:
step 601, generating a hybrid network structure according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes.
Step 602, for a target master node in the plurality of master nodes, determining a target slave node to be accessed by the target master node from the plurality of slave nodes.
Step 603, determining a first target packet in which a first routing node connected to the target master node is located.
In the embodiment of the present disclosure, a first routing node connected to a target master node may be obtained, a first packet in which the first routing node is located may be determined, and the first packet in which the first routing node is located may be used as a first target packet.
Step 604, a second destination packet is determined in which a second routing node to which the destination slave node is connected is located.
Similarly, a second routing node connected to the target slave node may be obtained, a second packet in which the second reason node is located may be determined, and the second packet in which the second routing node is located may be used as a second target packet.
In step 605, a third routing node connected to the second target packet is determined in the first target packet.
In the embodiment of the present disclosure, a routing node to which the second destination packet is connected may be determined in the first destination packet according to a hybrid network configuration, and the routing node may be regarded as the third routing node.
In step 606, a fourth routing node connected to the first destination packet is determined in the second destination packet.
Similarly, a routing node connected to the first destination packet may be determined in the second destination packet according to the hybrid network structure, and the routing node may be used as a fourth routing node.
Step 607, determining the target routing node according to the target master node, the target slave node, the third routing node and the fourth routing node.
In order to accurately determine the target routing node, it may be determined whether the target master node is directly connected to the third routing node, whether the target slave node is directly connected to the fourth routing node, that is, whether the target master node is connected to the third routing node via another routing node, and whether the target slave node is connected to the fourth routing node via another routing node.
As a possible implementation manner of the embodiment of the present disclosure, it may be determined, according to a device identifier between a master device node connected to a third routing node and a target master device node, whether at least one first associated routing node exists between the target master device node and the third routing node, where the first associated routing node is used to connect the target master device node and the third routing node, and at the same time, a location of a routing node connected to the target slave device node in a second target packet may be determined, and the location is compared with a location of a fourth routing node in the second target packet to determine whether at least one second associated routing node exists between the target slave device node and the fourth routing node, where the second associated routing node is used to connect the target slave device node and the fourth routing node.
As an example, in a case where there is at least one first associated routing node between the target master node and the third routing node, and there is at least one second associated routing node between the target slave node and the fourth routing node, the third routing node, the fourth routing node, the at least one first associated routing node, and the at least one second associated routing node are taken as the target routing nodes.
As another example, in the case where there is no at least one first associated routing node between the target master node and the third routing node, and there is at least one second associated routing node between the target slave node and the fourth routing node, the third routing node, the fourth routing node, and the at least one second associated routing node may be considered as target routing nodes.
As another example, in a case where there is at least one first associated routing node between the target master node and the third routing node, and there is no at least one second associated routing node between the target slave node and the fourth routing node, the third routing node, the fourth routing node, and the at least one first associated routing node are taken as the target routing node.
As another example, in the case where there is no at least one first associated routing node between the target master node and the third routing node, and there is no at least one second associated routing node between the target slave node and the fourth routing node, the third routing node and the fourth routing node are taken as the target routing nodes.
Step 608, determining an access path between the target master node and the target slave node according to the at least one target routing node.
It should be noted that the execution processes of steps 601 to 602 and step 608 may be implemented by any one of the embodiments of the present disclosure, and the embodiments of the present disclosure do not limit this, and are not described again.
In conclusion, a first target group where a first routing node connected with a target main equipment node is located is determined; determining a second target group in which a second routing node connected with the target slave device node is located; determining a third routing node in the first target packet that is connected to the second target packet; determining a fourth routing node in the second destination packet that is connected to the first destination packet; and determining the target routing node according to the target master device node, the target slave device node, the third routing node and the fourth routing node, so that the target routing node through which the target master device accesses the target slave device can be accurately determined.
In order to clearly illustrate the above embodiments, the description will now be made by way of example.
For example, as shown in fig. 7, the specific flow of the path generation method may be as follows:
1. m 0-m 15 are 16 master devices (master device nodes), S00-S73 are 32 routing nodes, under the condition that the 32 routing nodes are averagely divided into 8 groups, every 4 routing nodes form a group to form a small Mesh network, and the upper and lower 4 groups of Mesh networks have 8 groups of Mesh networks in total;
2. 8 groups of mesh networks are interconnected by adopting a Crossbar structure, 0-15 are 16 srams (slave equipment nodes) on the chip, and 16 masters can access each sram through a mixed NoC structure;
3. assuming that m0 wants to access the 11 th block sram, m0 passes through the routing node S00 of the 0 th group mesh network first and is routed to the node S03;
4. the routing node S03 is connected to a routing node S70 of the 7 th group of mesh networks through a Crossbar network;
5. then, the data passes through a routing node S71 in a 7 th group Mesh network and then reaches a node S72;
6. the shortest path finally routed by routing node S72 to the 11 th block sram, i.e., m0 accesses the 11 th block sram, is m0-S00-S03-S70-S71-S72-11.
According to the path generation method, a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and the target slave device node is determined, so that a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes, the situation that the communication transmission efficiency is low due to the network structure can be avoided, furthermore, in the process that the target master device accesses the target slave device, at least one target routing node can be selected from the hybrid network structure, and therefore according to the at least one target routing node, the optimal access path between the target master device node and the target slave device node can be effectively determined, and the communication transmission efficiency is improved.
In order to implement the above embodiments, the present disclosure also provides a path generation apparatus.
Fig. 8 is a schematic structural diagram of a path generation apparatus according to an exemplary embodiment.
As shown in fig. 8, the path generation apparatus 800 includes: a generation module 810, a first determination module 820, a second determination module 830, and a third determination module 840.
The generating module 810 is configured to generate a hybrid network structure according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes; a first determining module 820, configured to determine, for a target master node in the plurality of master nodes, a target slave node to be accessed by the target master node from the plurality of slave nodes; a second determining module 830, configured to determine at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; a third determining module 840, configured to determine an access path between the target master node and the target slave node according to the at least one target routing node.
As a possible implementation manner of the embodiment of the present disclosure, the generating module 810 is configured to: generating a first network structure according to the plurality of main equipment nodes and a first routing node matched with the plurality of main equipment nodes in the plurality of routing nodes; generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes; the first network fabric and the second network fabric are connected to generate a hybrid network fabric.
As a possible implementation manner of the embodiment of the present disclosure, the generating module 810 is further configured to: determining, from the plurality of routing nodes, a second routing node that matches each of the plurality of slave device nodes; grouping second routing nodes corresponding to the plurality of slave equipment nodes according to the number of the nodes to obtain a plurality of second groups; a second network structure is created from the second routing node in any of the second packets in the plurality of second packets and the slave node corresponding to the second routing node in any of the second packets.
As a possible implementation manner of the embodiment of the present disclosure, the generating module 810 is further configured to: acquiring a jth first routing node of an ith first packet in a plurality of first packets corresponding to a plurality of first network structures and an ith second routing node of a jth second packet in a plurality of second packets corresponding to a plurality of second network structures; connecting a jth first routing node of the ith first packet with an ith second routing node of the jth second packet to obtain a hybrid network structure; the number of the first groups and the number of the second groups are the same, the number of the first routing nodes in each first group is the same as the number of the first groups, and the number of the second routing nodes in each second group is the same as the number of the second groups.
As a possible implementation manner of the embodiment of the present disclosure, the generating module 810 is configured to: grouping the routing nodes to obtain a plurality of third groups, wherein each third group in the third groups comprises a plurality of routing nodes, the routing nodes in each third group are connected by adopting a third network structure, and the third groups are connected by adopting a fourth network structure; connecting the plurality of master device nodes and the plurality of slave device nodes through the third network fabric and the fourth network fabric to generate a hybrid network fabric.
As a possible implementation manner of the embodiment of the present disclosure, the second determining module 830 is configured to: determining a first target group in which a first routing node connected with a target main equipment node is positioned; determining a second target group in which a second routing node connected with the target slave device node is located; determining a third routing node connected with the second target packet in the first target packet; determining a fourth routing node connected with the first target packet in the second target packet; and determining the target routing node according to the target master equipment node, the target slave equipment node, the third routing node and the fourth routing node.
As a possible implementation manner of the embodiment of the present disclosure, the second determining module 830 is further configured to: determining whether at least one first associated routing node exists between the target main equipment node and a third routing node, wherein the first associated routing node is used for connecting the target main equipment node and the third routing node; determining whether at least one second associated routing node exists between the target slave device node and a fourth routing node, wherein the second associated routing node is used for connecting the target slave device node and the fourth routing node; and under the condition that at least one first associated routing node exists between the target master equipment node and the third routing node and at least one second associated routing node exists between the target slave equipment node and the fourth routing node, taking the third routing node, the fourth routing node, the at least one first associated routing node and the at least one second associated routing node as target routing nodes.
As a possible implementation manner of the embodiment of the present disclosure, the path generating apparatus 800 further includes: the device comprises a fourth determination module, a fifth determination module and a sixth determination module.
The fourth determining module is configured to, in a case that at least one first associated routing node does not exist between the target master device node and the third routing node, and at least one second associated routing node exists between the target slave device node and the fourth routing node, take the third routing node, the fourth routing node, and the at least one second associated routing node as the target routing node; a fifth determining module, configured to, when at least one first associated routing node exists between the target master device node and the third routing node and at least one second associated routing node does not exist between the target slave device node and the fourth routing node, take the third routing node, the fourth routing node, and the at least one first associated routing node as the target routing node; a sixth determining module, configured to take the third routing node and the fourth routing node as target routing nodes when at least one first associated routing node does not exist between the target master device node and the third routing node and at least one second associated routing node does not exist between the target slave device node and the fourth routing node.
The path generation device of the embodiment of the present disclosure generates a hybrid network structure according to a plurality of master device nodes, a plurality of slave device nodes, and a plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and the target slave device node is determined, so that a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes, the situation that the communication transmission efficiency is low due to the network structure can be avoided, furthermore, in the process that the target master device accesses the target slave device, at least one target routing node can be selected from the hybrid network structure, and therefore according to the at least one target routing node, the optimal access path between the target master device node and the target slave device node can be effectively determined, and the communication transmission efficiency is improved.
In order to implement the above embodiments, the present disclosure also proposes an electronic device, as shown in fig. 9, where fig. 9 is a block diagram of an electronic device for path generation shown according to an exemplary embodiment. As shown in fig. 9, the electronic device 900 may include:
a memory 910 and a processor 920, and a bus 930 connecting different components (including the memory 910 and the processor 920), wherein the memory 910 stores a computer program, and when the processor 920 executes the program, the path generation method according to the embodiment of the disclosure is implemented.
Bus 930 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, industry Standard Architecture (ISA) bus, micro-channel architecture (MAC) bus, enhanced ISA bus, video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.
Electronic device 900 typically includes a variety of computer-readable media. Such media may be any available media that is accessible by electronic device 900 and includes both volatile and nonvolatile media, removable and non-removable media.
The memory 910 may also include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM) 940 and/or cache memory 950. The electronic device 900 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 960 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 9, commonly referred to as a "hard drive"). Although not shown in FIG. 9, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 930 by one or more data media interfaces. Memory 910 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the disclosure.
A program/utility 980 having a set (at least one) of program modules 970, which may be stored for example in memory 910, such program modules 970 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof may comprise an implementation of a network environment. The program modules 970 generally perform the functions and/or methodologies of the embodiments described in this disclosure.
The electronic device 900 may also communicate with one or more external devices 990 (e.g., keyboard, pointing device, display 991, etc.), with one or more devices that enable a user to interact with the electronic device 900, and/or with any devices (e.g., network card, modem, etc.) that enable the electronic device 900 to communicate with one or more other computing devices. Such communication may occur via an input/output (I/O) interface 992. Also, the electronic device 900 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 993. As shown in FIG. 9, the network adapter 993 communicates with the other modules of the electronic device 900 via the bus 930. It should be appreciated that although not shown in FIG. 9, other hardware and/or software modules may be used in conjunction with electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
The processor 920 performs various functional applications and data processing by executing programs stored in the memory 910.
It should be noted that, for the implementation process and the technical principle of the electronic device of the embodiment, reference is made to the foregoing explanation of the path generation method of the embodiment of the present disclosure, and details are not described here again.
According to the electronic device provided by the embodiment of the disclosure, a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes; aiming at a target master device node in a plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from a plurality of slave device nodes; determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node; according to at least one target routing node, an access path between a target master device node and the target slave device node is determined, so that a hybrid network structure is generated according to a plurality of master device nodes, a plurality of slave device nodes and a plurality of routing nodes, the situation that the communication transmission efficiency is low due to the network structure can be avoided, furthermore, in the process that the target master device accesses the target slave device, at least one target routing node can be selected from the hybrid network structure, and therefore according to the at least one target routing node, the optimal access path between the target master device node and the target slave device node can be effectively determined, and the communication transmission efficiency is improved.
In order to implement the above embodiments, the embodiments of the present disclosure also provide a computer-readable storage medium.
Wherein the instructions in the computer readable storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the path generation method as previously described.
To implement the above embodiments, the present disclosure also provides a computer program product, which, when executed by a processor of an electronic device, enables the electronic device to perform the path generation method as described above.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims (14)

1. A method for generating a path, comprising:
generating a hybrid network structure according to the plurality of master device nodes, the plurality of slave device nodes and the plurality of routing nodes;
for a target master device node in the plurality of master device nodes, determining a target slave device node to be accessed by the target master device node from the plurality of slave device nodes;
determining at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node;
and determining an access path between the target master device node and the target slave device node according to the at least one target routing node.
2. The method of claim 1, wherein generating a hybrid network structure from the plurality of master nodes, the plurality of slave nodes, and the plurality of routing nodes comprises:
generating a first network structure according to the plurality of main equipment nodes and a first routing node matched with the plurality of main equipment nodes in the plurality of routing nodes;
generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes;
connecting the first network fabric and the second network fabric to generate a hybrid network fabric.
3. The method of claim 2, wherein generating the first network fabric based on the plurality of master nodes and a first routing node of the plurality of routing nodes that matches the plurality of master nodes comprises:
the plurality of main equipment nodes can be sequenced according to the equipment identification information so as to obtain the sequencing sequence of the plurality of main equipment nodes;
determining a first routing node matched with each main equipment node in the plurality of main equipment nodes from the plurality of routing nodes according to the arrangement sequence of the plurality of main equipment nodes;
grouping first routing nodes corresponding to the plurality of main equipment nodes according to the set number of the nodes to obtain a plurality of first groups;
and creating a first network structure according to the first routing node in any one of the plurality of first groups and the main equipment node corresponding to the first routing node in any one of the plurality of first groups.
4. The method of claim 3, wherein generating a second network structure based on the plurality of slave device nodes and a second routing node of the plurality of routing nodes that matches the plurality of slave device nodes comprises:
determining, from the plurality of routing nodes, a second routing node that matches each of the plurality of slave device nodes;
grouping second routing nodes corresponding to the plurality of slave equipment nodes according to the number of the nodes to obtain a plurality of second groups;
and creating a second network structure according to the second routing node in any second packet in the second packets and the slave device node corresponding to the second routing node in any second packet.
5. The method of claim 4, wherein the connecting the first network fabric and the second network fabric to generate a hybrid network fabric comprises:
acquiring a jth first routing node of an ith first packet in a plurality of first packets corresponding to the first network structures and an ith second routing node of a jth second packet in a plurality of second packets corresponding to the second network structures;
connecting the jth first routing node of the ith first packet with the ith second routing node of the jth second packet to obtain a hybrid network structure;
the number of the first packets and the number of the second packets are the same, the number of the first routing nodes in each first packet is the same as the number of the first packets, and the number of the second routing nodes in each second packet is the same as the number of the second packets.
6. The method of claim 1, wherein generating a hybrid network structure from the plurality of master nodes, the plurality of slave nodes, and the plurality of routing nodes comprises:
grouping the routing nodes to obtain a plurality of third groups, wherein each third group in the third groups comprises a plurality of routing nodes, the routing nodes in each third group are connected by adopting a third network structure, and the third groups are connected by adopting a fourth network structure;
connecting the plurality of master device nodes and the plurality of slave device nodes through the third network fabric and the fourth network fabric to generate the hybrid network fabric.
7. The method of claim 1, wherein determining at least one target routing node from the hybrid network structure based on the target master node and the target slave node comprises:
determining a first target group in which a first routing node connected with the target main equipment node is positioned;
determining a second target packet in which a second routing node connected to the target slave node is located;
determining, in the first destination packet, a third routing node connected to the second destination packet;
determining a fourth routing node in the second destination packet that is connected to the first destination packet;
and determining a target routing node according to the target master device node, the target slave device node, the third routing node and the fourth routing node.
8. The method of claim 7, wherein determining a target routing node from the target master node, the target slave node, a third routing node, and the fourth routing node comprises:
determining whether at least one first associated routing node exists between the target master node and the third routing node, wherein the first associated routing node is used for connecting the target master node and the third routing node;
determining whether at least one second associated routing node exists between the target slave device node and the fourth routing node, wherein the second associated routing node is used for connecting the target slave device node and the fourth routing node;
and taking the third routing node, the fourth routing node, the at least one first associated routing node and the at least one second associated routing node as target routing nodes under the condition that at least one first associated routing node exists between the target master device node and the third routing node and at least one second associated routing node exists between the target slave device node and the fourth routing node.
9. The method of claim 8, further comprising:
taking the third routing node, the fourth routing node, and the at least one second associated routing node as target routing nodes if at least one first associated routing node does not exist between the target master device node and the third routing node and at least one second associated routing node exists between the target slave device node and the fourth routing node;
taking the third routing node, the fourth routing node and the at least one first associated routing node as target routing nodes if at least one first associated routing node exists between the target master device node and the third routing node and at least one second associated routing node does not exist between the target slave device node and the fourth routing node;
and taking the third routing node and the fourth routing node as target routing nodes under the condition that at least one first associated routing node does not exist between the target master device node and the third routing node and at least one second associated routing node does not exist between the target slave device node and the fourth routing node.
10. A path generation apparatus, comprising:
the generation module is used for generating a hybrid network structure according to the plurality of master equipment nodes, the plurality of slave equipment nodes and the plurality of routing nodes;
a first determining module, configured to determine, for a target master node in the plurality of master nodes, a target slave node to be accessed by the target master node from the plurality of slave nodes;
a second determining module, configured to determine at least one target routing node from the hybrid network structure according to the target master device node and the target slave device node;
a third determining module, configured to determine, according to the at least one target routing node, an access path between the target master device node and the target slave device node.
11. The apparatus of claim 10, wherein the generating module is configured to:
generating a first network structure according to the plurality of main equipment nodes and a first routing node matched with the plurality of main equipment nodes in the plurality of routing nodes;
generating a second network structure according to the plurality of slave device nodes and a second routing node matched with the plurality of slave device nodes in the plurality of routing nodes;
connecting the first network fabric and the second network fabric to generate a hybrid network fabric.
12. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the path generation method of any of claims 1-9.
13. A computer-readable storage medium whose instructions, when executed by a processor of an electronic device, enable the electronic device to perform the path generation method of any of claims 1-9.
14. A computer program product comprising a computer program which, when executed by a processor of an electronic device, enables the electronic device to perform the path generation method of any of claims 1-9.
CN202211471812.5A 2022-11-23 2022-11-23 Path generation method and device and electronic equipment Pending CN115865784A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211471812.5A CN115865784A (en) 2022-11-23 2022-11-23 Path generation method and device and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211471812.5A CN115865784A (en) 2022-11-23 2022-11-23 Path generation method and device and electronic equipment

Publications (1)

Publication Number Publication Date
CN115865784A true CN115865784A (en) 2023-03-28

Family

ID=85665206

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211471812.5A Pending CN115865784A (en) 2022-11-23 2022-11-23 Path generation method and device and electronic equipment

Country Status (1)

Country Link
CN (1) CN115865784A (en)

Similar Documents

Publication Publication Date Title
Ajima et al. The tofu interconnect d
US11934883B2 (en) Computer cluster arrangement for processing a computation task and method for operation thereof
US10027776B2 (en) Distributed queue pair state on a host channel adapter
CN107710238B (en) Deep neural network processing on hardware accelerators with stack memory
JP4560435B2 (en) High performance computing system and high performance computing method
JP4833965B2 (en) System and method for cluster management based on HPC architecture
JP4986844B2 (en) System and method for detecting and managing HPC node failures
JP2005310139A (en) System and method for executing topology-aware job scheduling and backfilling in hpc environment
JP2006146864A (en) Scheduling in high-performance computing (hpc) system
CN109062613B (en) Multi-core interconnection secondary cache access verification method
De Prisco et al. A dynamic view-oriented group communication service
US11409684B2 (en) Processing accelerator architectures
CN114896182B (en) Storage device, method, electronic apparatus, and storage medium
CN108304261B (en) Job scheduling method and device based on 6D-Torus network
CN110727611B (en) Configurable consistency verification system with state monitoring function
Balkan et al. An area-efficient high-throughput hybrid interconnection network for single-chip parallel processing
CN115865784A (en) Path generation method and device and electronic equipment
CN112579293B (en) Comprehensive verification method of distributed computing system
Rupprecht Exploiting in-network processing for big data management
AT&T
CN114363988A (en) Clustering method and device and electronic equipment
US20060143357A1 (en) Multiple cell computer systems and methods
Avresky et al. Performance Modeling of ServerNet™ SAN Topologies
CN110347343A (en) Data managing method and device
CN112765925B (en) Interconnected circuit system, verification system and method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination