Construction method, device and equipment of reconstruction network based on system-on-chip, and medium

Abstract

The invention provides a construction method and device of a reconstruction network based on a system-on-chip, equipment and a medium. The method comprises the following steps: acquiring an initial network topology, wherein the topology comprises a standby link and a common link which are respectively connected between a current router and an adjacent next router of a pre-constructed system-on-chip; the normal state of the standby link is a non-enabled state; in the process of using the common link to carry out data transmission, under the condition that the common link is detected to be congested, starting a corresponding standby link to carry out data transmission, updating the link state of the router of the initial network topology, and constructing a current network topology comprising the standby link in the starting state and the common link in the using state; and when it is detected that the congestion of the common link disappears and the link is idle, closing the corresponding standby link, updating the link state of the router of the initial network topology, and constructing the current network topology comprising the common link in the use state. Thus, the network can be dynamically reconstructed.

Description

Technical Field

[0001] The present invention relates to the technical field of network reconstruction, and particularly to a method, apparatus, device, and medium for constructing a reconstructed network based on a system on wafer. Background Art

[0002] With the rapid development of integrated circuit technology, a system on wafer (SoW, also known as a wafer-level chip) has gradually become an important platform for high-performance computing and large-scale data processing. A system on wafer integrates a large number of computing units and communication links on a single wafer to achieve extremely high computing density and data transmission efficiency. However, with the expansion of chip scale and the diversification of application scenarios, the network design of a system on wafer faces many challenges.

[0003] In related system-on-wafer designs, the network topology usually adopts a fixed design scheme. These fixed topologies need to consider the communication requirements in the worst case during design to ensure that the system can operate normally under any circumstances.

[0004] Therefore, in the actual use process, how to dynamically reconstruct the network when congestion occurs in the network of a system on wafer has become an urgent technical problem to be solved. Summary of the Invention

[0005] This application provides a method, apparatus, device, and medium for constructing a reconstructed network based on a system on wafer.

[0006] This application provides a method for constructing a reconstructed network based on a system on wafer, including: Obtain an initial network topology; the initial network topology includes: standby links and common links respectively connected between the current router and the adjacent next router of a pre-constructed system on wafer; the normal state of the standby links is a non-enabled state; During the process of data transmission using the common links, when it is detected that congestion occurs in the common links, enable the corresponding standby links for data transmission, and update the link status of the routers in the initial network topology to construct the current network topology including the enabled standby links and the used common links; When it is detected that the congestion of the common links disappears and the links are idle, close the corresponding standby links, and update the link status of the routers in the initial network topology to construct the current network topology including the used common links.

[0007] Further, during the process of data transmission using the common links, when it is detected that congestion occurs in the common links, enabling the corresponding standby links for data transmission includes: In the case of detecting congestion in the common link, determine the real-time congestion situation; In the case of determining that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enable the corresponding backup link.

[0008] Further, the real-time congestion situation includes the current congestion degree: The step of automatically and dynamically enabling the corresponding backup link in the case of determining that the real-time congestion situation meets the condition of increasing congestion includes: When the current congestion degree is greater than the congestion threshold, enable the backup link for data transmission.

[0009] Further, the number of backup links is more than 1; The step of enabling the backup link for data transmission when the current congestion degree is greater than the congestion threshold includes: When the current congestion degree is greater than the lowest congestion threshold, enable 1 backup link for data transmission; When the current congestion degree is greater than the highest congestion threshold, enable more than 2 backup links for data transmission simultaneously or successively; the lowest congestion threshold is less than the highest congestion threshold.

[0010] Further, the number of backup links is more than 1; the real-time congestion situation includes the changing degree of the current congestion; The step of automatically and dynamically enabling the corresponding backup link in the case of determining that the real-time congestion situation meets the condition of increasing congestion includes: When the current congestion degree is greater than the lowest congestion threshold, enable the first backup link for data transmission; When the changing degree of the current congestion reaches the condition of gradual increase, gradually enable the remaining backup links after the first backup link for data transmission.

[0011] Further, the real-time congestion situation includes the current congestion degree: The step of automatically and dynamically enabling the corresponding backup link in the case of determining that the real-time congestion situation meets the condition of increasing congestion includes: In the case of determining that the real-time congestion situation meets the condition of increasing congestion, based on the current congestion degree, determine the number of enabled backup links corresponding to the current congestion degree according to the corresponding relationship established between the congestion degree and the number of enabled backup links; Enable the corresponding backup links with the determined number of enabled links for data transmission.

[0012] Further, the number of the standby links is more than one; in the case that it is detected that the congestion of the common link disappears and the link is idle, closing the corresponding standby link includes: In the case that it is detected that the common link is congested, determining the real-time congestion situation; In the case that it is determined that the real-time congestion situation meets the congestion reduction condition, maintaining the currently enabled standby link for data transmission until it is detected that the congestion of the common link disappears and the link is idle, and then closing the corresponding standby link, or gradually closing the currently enabled standby link until it is detected that the congestion of the common link disappears and the link is idle, and then closing all the enabled standby links.

[0013] Further, the following method is adopted to determine that the real-time congestion situation meets the congestion reduction condition: The real-time congestion situation includes the change degree of the current congestion. Correspondingly, the change degree of the current congestion reaches the gradually decreasing condition; Or, The real-time congestion situation includes the current congestion degree; correspondingly, the current congestion degree is lower than the lowest congestion threshold.

[0014] Further, enabling the corresponding standby link for data transmission includes: If receiving a data packet for each data transmission, randomly selecting either the common link or the standby link from the common link and the standby link for data transmission; Or, If receiving a data packet for each data transmission, preferentially using the common link for transmission, and when the common link reaches the preset stop transmission threshold, then selecting the standby link for data transmission; Or, If receiving more than two data packets for each data transmission, separating the more than two data packets into independent sub-packets; and respectively using the standby link and the common link to transmit the independent sub-packets.

[0015] Further, the method further includes: During the process of using the common link for data transmission, adopting the following method to detect whether the common link is congested: During the process of using the common link for data transmission, determining the current buffer data volume of the buffer of the system on the chip; If the current buffer data volume reaches the congestion condition, it is detected that the common link is congested; If the current buffer data volume does not reach the congestion condition, it is detected that the common link is not congested.

[0016] The present application provides a device for constructing a reconfigurable network based on a system-on-chip, including: A status determination module, configured to obtain an initial network topology; the initial network topology includes: a standby link and a common link respectively connected between a current router and an adjacent next router of a pre-constructed system-on-chip; the normal state of the standby link is a non-enabled state; An enabling detection module, configured to, during the process of using the common link for data transmission, in the case of detecting congestion of the common link, enable a corresponding standby link for data transmission, and update the link status of the routers in the initial network topology, to construct a current network topology including the standby link in an enabled state and the common link in a used state; A closing detection module, configured to, in the case of detecting that the congestion of the common link disappears and the link is idle, close the corresponding standby link, and update the link status of the routers in the initial network topology, to construct a current network topology including the common link in a used state.

[0017] The present application provides an electronic device, including one or more processors, configured to implement the method described in any one of the above.

[0018] The present application provides a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the method described in any one of the above is implemented.

[0019] The present application provides a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the method described in any one of the above is implemented.

[0020] In some embodiments, for the method for constructing a reconfigurable network based on a system-on-chip of the present application, a standby link with a normal state of non-enabled is added to the initial network topology. In this way, when not in use, no energy is consumed. For congestion at a specific local location, the corresponding standby link at the local location is enabled, without affecting the global situation. In this way, only enabling the local standby link can also reduce energy consumption. At the same time, when the common link between the current router and the adjacent next router is congested, a local standby link is added for enabling. In this way, when congestion occurs in the network of the system-on-chip, multiple links such as the common link and the standby link are used for data transmission, which can not only realize dynamic reconfiguration of the network, but also improve the transmission efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The figure shows a flowchart of the method for constructing a reconfigurable network based on a system-on-chip provided by an embodiment of the present application; Figure 2 As shown in Figure 1Schematic diagram of the network topology structure of the construction method of the reconfiguration network based on the system-on-chip shown; Figure 3 As shown; Figure 1 Schematic diagram of the network topology structure of the reconfigured link when link congestion occurs in the construction method of the reconfiguration network based on the system-on-chip shown; Figure 4 As shown; Figure 1 Specific flowchart of the construction method of the reconfiguration network based on the system-on-chip shown; Figure 5 Schematic diagram of the structure of the construction device of the reconfiguration network based on the system-on-chip provided by the embodiment of the present application; Figure 6 Schematic diagram of the structure of the electronic device provided by the embodiment of the present application. Detailed implementation manners

[0022] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. On the contrary, they are merely examples of devices and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0023] It should be noted that: in other embodiments, the steps of the corresponding methods are not necessarily executed in the order shown and described in this specification. In some other embodiments, the steps included in the method may be more or less than those described in this specification. In addition, a single step described in this specification may be decomposed into multiple steps for description in other embodiments; and multiple steps described in this specification may also be combined into a single step for description in other embodiments.

[0024] In order to solve the above technical problem that the network of the system-on-chip is congested and the dynamic reconfiguration of the network becomes an urgent problem to be solved, the embodiment of the present application provides a construction method of a reconfiguration network based on the system-on-chip, and adds standby links that are normally in a non-enabled state to the initial network topology. In this way, when not in use, energy will not be consumed (abbreviated as energy consumption). For specific local congestion, the corresponding standby links at the local position are enabled, without affecting the global situation. In this way, only enabling the local standby links can also reduce energy consumption. At the same time, when the common link between the current router and the adjacent next router is congested, local standby links are added for enabling. In this way, when the network of the system-on-chip is congested, multiple links such as common links and standby links are used for data transmission, which can not only realize the dynamic reconfiguration of the network, but also improve the transmission efficiency.

[0025] Figure 1 The figure shows a schematic flow chart of a method for constructing a reconfigurable network based on a system-on-chip according to an embodiment of the present application.

[0026] As Figure 1 shown, the method for constructing the reconfigurable network based on the system-on-chip may include but is not limited to the following steps 110 to 130: Step 110, obtaining an initial network topology; the initial network topology includes: a standby link and a common link respectively connected between the current router and the adjacent next router of a pre-constructed system-on-chip; the normal state of the standby link is a non-enabled state.

[0027] Among them, the network physical links of the above system-on-chip are manufactured according to the most commonly used 2D Mesh (Two-Dimensional Mesh) topology structure, and standby links, that is, metal signal lines, are designed between routers. Among them, 2D Mesh is the most commonly used on-chip interconnection network topology structure. The system-on-chip can also be called a system-on-wafer. This system directly integrates multiple functional components on a single wafer, and through advanced packaging technology, realizes high integration and efficient interconnection to form a complete system.

[0028] In addition, the network topology is initialized, which is called the initial network topology. Among them, the initialization may include initializing the standby link of the network topology to be non-enabled (also called the normal state), that is, not powered on. In this way, power consumption and heat generation are not brought. The initialization may also include initializing the local congestion awareness of the router.

[0029] As Figure 2 shown, a router (abbreviated as R) will automatically select and set a route according to the situation of the channel, and send signals in the best path in the front-to-back order. There can be many routers R in the network. As Figure 2 shown, the exemplary routers R include router R1, router R2, router R3, router R4, router R5, router R6, router R7, router R8 and router R9. Figure 2 The solid lines in it represent common links. The common link is always powered on and in an enabled state, which is used to indicate that the link is continuously used after the system-on-chip is powered on. One or more common links can be set in the embodiment of the present application.

[0030] The standby link in this article is normally in a non-enabled state, that is, represented by Figure 2 the dotted line. When the common link is congested, the temporary state of the standby link is an enabled state, and the standby link represented by the dotted line is changed to the standby link represented by the solid line, so that the standby link becomes a part of the current network topology. Exemplarily, Figure 2 for the router R7, the dotted line passing through router R8 to router R9 becomesFigure 3 The solid line from router R7 to router R8 and then to router R9. In this way, users only need to initially arrange the backup link once. During the usage process, they can re-enable the backup link to reconstruct the current network architecture without having to re-arrange each link again.

[0031] Step 120, during the process of data transmission using the common link, when it is detected that the common link is congested, enable the corresponding backup link for data transmission, and update the link status of the routers in the initial network topology to construct the current network topology that includes the enabled backup link and the in-use common link.

[0032] For the above step 120, the routers during network operation perform local congestion awareness, such as detecting congestion in the common link between the current router and the next adjacent router. That is, when congestion is currently detected, the connected backup link is automatically and dynamically enabled through the routing algorithm. Since only a local area is enabled, it will not affect other paths. The routing algorithm is used to automatically and dynamically select and use the links of this route.

[0033] Step 130, when it is detected that the congestion of the common link has disappeared and the link is idle, close the corresponding backup link, and update the link status of the routers in the initial network topology to construct the current network topology that includes the in-use common link. In this way, when it is detected that the congestion has disappeared and the link is idle, the backup link will be closed to reduce power consumption.

[0034] For the construction of the current network topology in steps 120 and 130 in this article, it can also be called reconstruction. This reconstruction behavior can be dynamic. During the real-time operation of the current application, that is, adding or reducing communication paths for new applications on the current available network resources without affecting the normal transmission of other data packets. Therefore, its reconstruction and execution can be carried out crosswise.

[0035] In the dynamic reconstruction method of the network of the system-on-chip in the related technology, it is directly transferred from the input to the output through the internal logic of the router. Specifically, the first common link is from the input router A, through routers B and C, and then to the output router D. It is directly transferred from the input router A to the output router D through the internal logic of the router, and a second common link from router A to the output router D is added.

[0036] However, in the actual use of the related technology, the system-on-chip will detect the first common link and the second common link, consuming energy continuously, resulting in relatively large energy consumption. At the same time, the links of the system-on-chip are fixed and the flexibility is low.

[0037] Compared with the related art, in the embodiments of the present application, when the common link between the current router and the next adjacent router is congested, the local standby link is enabled. In this way, the energy consumption is small. At the same time, the standby link is enabled at the congested location locally. And, when the congestion disappears at a certain location and the link is idle, the corresponding standby link is closed. It can be seen that the corresponding opened and closed links make the network usage more flexible.

[0038] Combined with Figure 1 As shown, after the standby link is enabled, the link state in the router needs to be modified to inform the routing algorithm module that the link can be used for routing selection.

[0039] On the one hand, the following method is adopted to determine to enable the corresponding standby link for data transmission, which specifically includes the following Step 1 and Step 2: Step 1, in the case of detecting that the common link is congested, determine the real-time congestion situation.

[0040] Step 2, in the case of determining that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enable the corresponding standby link. In this way, according to the real-time congestion situation, it can be adaptively and dynamically determined whether to enable the standby link.

[0041] The condition of increasing congestion in this article is used to indicate that the congestion is getting more and more serious, that is, the amount of congested data is increasing. The condition of decreasing congestion in this article is used to indicate that the congestion is getting less and less, that is, the amount of congested data is decreasing.

[0042] Based on this, the above Step 2 adopts the following method to automatically and dynamically enable the corresponding standby link: the above real-time congestion situation may include, but is not limited to, the current congestion degree. Correspondingly, when the current congestion degree is greater than the congestion threshold, the standby link is enabled for data transmission. Among them, the above current congestion degree being greater than the congestion threshold is used to indicate that the common link is congested.

[0043] In this regard, in the first optional example of enabling the standby link for data transmission when the current congestion degree is greater than the congestion threshold, the number of standby links is more than 1. First, when the current congestion degree is greater than the lowest congestion threshold, 1 standby link is enabled for data transmission. And, secondly, when the current congestion degree is greater than the highest congestion threshold, 2 or more standby links are enabled for data transmission simultaneously or successively; the lowest congestion threshold is less than the highest congestion threshold. The "successively" here can be the successive order of adjacent consecutive times, or the successive order of interval times.

[0044] The congestion threshold in this article can be one, or two or more (including two). Multiple congestion thresholds can be, but are not limited to, the lowest congestion threshold and / or the highest congestion threshold. The lowest congestion threshold is used to indicate a relatively low degree of congestion. When the current congestion degree is greater than the lowest congestion threshold, it indicates the initial stage of congestion, which means that the user is relatively conservative. As long as congestion occurs, the backup link is enabled. In this way, the smoothness of the link can be ensured. The highest congestion threshold is used to indicate a relatively high degree of congestion. When the current congestion degree is greater than the highest congestion threshold, it indicates that congestion has occurred for a period of time and the congestion is continuously increasing, which means that one backup link may not be enough. Therefore, two or more backup links are enabled simultaneously or successively for data transmission. The highest congestion threshold is greater than the lowest congestion threshold, indicating that the user is relatively aggressive and the network needs to be congested enough before enabling two or more backup links for data transmission.

[0045] In addition, in the second alternative example of enabling the backup link for data transmission when the current congestion degree is greater than the congestion threshold, the number of backup links is one or more; the real-time congestion situation includes the change degree of the current congestion. First, when the current congestion degree is greater than the lowest congestion threshold, the first backup link is enabled for data transmission; and second, when the change degree of the current congestion reaches the condition of gradually increasing, the remaining backup links after the first backup link are gradually enabled for data transmission. Among them, "gradually" means at intervals of a preset time or at intervals of a preset number of data packets. Herein, the "preset time" and / or "preset number" in this article are used to represent the control rhythm of enabling the backup link. The longer the preset time and / or the larger the preset number, the slower the speed of gradually enabling the remaining backup links after the first backup link. The smaller the preset time and / or the smaller the preset number, the faster the speed of gradually enabling the remaining backup links after the first backup link.

[0046] Furthermore, in the third alternative example of enabling the backup link for data transmission when the current congestion degree is greater than the congestion threshold, the real-time congestion situation includes the current congestion degree. First, when it is determined that the real-time congestion situation meets the condition of increasing congestion, based on the current congestion degree, according to the corresponding relationship established between the congestion degree and the number of enabled backup links, the number of enabled backup links corresponding to the current congestion degree is determined. And second, the corresponding relationship established between the above-mentioned congestion degree and the number of enabled backup links is used to indicate that the greater the congestion degree, the more the number of startups, so as to transmit congestion data more timely, make the link quickly become smooth, and the faster the data transmission speed.

[0047] It should be noted that the corresponding relationship established between the above congestion degree and the number of enabled backup links may include, but is not limited to, the corresponding relationship where the percentage reached by the above congestion degree is the same as the percentage of the total number of corresponding backup links and is rounded up, or the percentage reached by the above congestion degree is the same as the percentage of the total number of corresponding backup links and is rounded down. Either rounding up or rounding down is acceptable in this article. Of course, according to the actual situation, the higher the degree of congestion, the more it can be rounded up.

[0048] For example, when the channel condition is fixed and the congestion degree reaches 50%, the number of enabled backup links corresponding thereto can be 50% of the total number of backup links, and either rounding up or rounding down is acceptable. For another example, when the congestion degree reaches 60%, the number of enabled backup links corresponding thereto can be 60% of the total number of backup links, and either rounding up or rounding down is acceptable.

[0049] It should also be noted that the corresponding relationship established between the above congestion degree and the number of enabled backup links may include, but is not limited to, the one-to-one correspondence between the percentage reached by the above congestion degree and the number of enabled corresponding backup links. The higher the congestion degree, the more the number of enabled backup links. As the congestion degree grows faster, the number of enabled backup links is also more and faster.

[0050] For example, when the channel condition is fixed and the congestion degree reaches 50%, the number of enabled backup links corresponding thereto can be 1. For another example, when the congestion degree reaches 60%, the number of enabled backup links corresponding thereto can be 2. For yet another example, when the congestion degree reaches 65%, the number of enabled backup links corresponding thereto can be 3. In this way, the backup links corresponding to the number of enabled ones are used for data transmission.

[0051] On the other hand, when the number of backup links is more than 1, in the case where it is detected that the congestion of the common link disappears and the link is idle, the following optional method is adopted to close the corresponding backup link, including: ①. In the case where it is detected that the common link is congested, determine the real-time congestion situation.

[0052] ②. In the case where it is determined that the real-time congestion situation meets the condition of congestion reduction, maintain the currently enabled backup links for data transmission until it is detected that the congestion of the common link disappears and the link is idle, then close the corresponding backup link, or gradually close the currently enabled backup links until it is detected that the congestion of the common link disappears and the link is idle, and then close all the enabled backup links to construct the current network including the common link.

[0053] Among them, the following any one method is adopted to determine that the real-time congestion situation meets the condition of congestion reduction: In a first alternative manner, the real-time congestion situation includes the degree of change of the current congestion. Correspondingly, the degree of change of the current congestion reaches a gradually decreasing condition.

[0054] In a second alternative manner, the real-time congestion situation includes the current congestion degree; correspondingly, the current congestion degree is lower than the lowest congestion threshold.

[0055] Based on this, an example scheme for closing the enabled backup link is as follows: Example 1: When the degree of change of the current congestion reaches the gradually decreasing condition, maintain the currently enabled backup link for data transmission until it is detected that the congestion of the common link disappears and the link is idle, and then close the corresponding backup link.

[0056] Example 2: When the degree of change of the current congestion reaches the gradually decreasing condition, gradually close the currently enabled backup link until it is detected that the congestion of the common link disappears and the link is idle, and then close all enabled backup links.

[0057] Example 3: When the current congestion degree is lower than the lowest congestion threshold, maintain the currently enabled backup link for data transmission until it is detected that the congestion of the common link disappears and the link is idle, and then close the corresponding backup link.

[0058] Example 4: When the current congestion degree is lower than the lowest congestion threshold, gradually close the currently enabled backup link until it is detected that the congestion of the common link disappears and the link is idle, and then close all enabled backup links.

[0059] Combined with Figure 1 As shown, any of the following methods can be used to implement data transmission using the corresponding enabled backup link in step 120 above: In a first alternative manner, if a data packet for each data transmission is received, randomly select either the common link or the backup link from the common link and the backup link for data transmission.

[0060] In a second alternative manner, if a data packet for each data transmission is received, give priority to using the common link for transmission. When the common link reaches the preset stop transmission threshold, then select the backup link for data transmission.

[0061] The above preset stop transmission threshold is used to represent the threshold of link data congestion. This threshold can be set according to user requirements.

[0062] In a third alternative manner, if more than two data packets for each data transmission are received, divide the more than two data packets into independent sub-packets; and use the backup link and the common link to transmit the independent sub-packets respectively.

[0063] Among them, this article can adopt at least one of the following examples to divide more than two data packets into independent sub - data packets: In the first example, if the number of more than two data packets is a multiple of 2, the more than two data packets are equally divided into two parts of sub - data packets. Specifically, the data packets for the backup link can be distributed to each backup link for transmission respectively.

[0064] In the second example, if the number of more than two data packets is not a multiple of 2, the more than two data packets are unequally divided into two parts of sub - data packets. Specifically, the data packets for the backup link are larger than the other part for the normal link, and the data packets for the backup link can be distributed to each backup link for transmission respectively.

[0065] In the third example, the number of sub - data packets into which more than two data packets are divided is the same as the total number of backup links and normal links. And each divided data packet is transmitted using the backup link and the normal link for data transmission respectively.

[0066] Figure 4 The following is a flowchart for congestion detection after enabling the backup link in the embodiment of the present invention.

[0067] As Figure 4 shown, the method for constructing a reconfigurable network based on a system - on - chip in this article can also include the following steps 111 to 113 during the process of data transmission using the normal link to detect whether the normal link is congested: Step 111: During the process of data transmission using the normal link, determine the current buffer data volume of the buffer of the system - on - chip. This current buffer data volume can be the real - time buffer data volume.

[0068] Among them, the buffer in this article is used to temporarily store the data waiting for transmission. When the data in the buffer is transmitted, the buffer will increase as packets are received, and will notify the routing algorithm module of the remaining buffer capacity. At the same time, the data packets in the buffer will decrease as data is transmitted, which changes in real - time as data is transmitted.

[0069] The current buffer data volume of the buffer is used to illustrate whether the current buffer data volume is excessive and whether it reaches the congestion condition. The current buffer data volume is negatively correlated with the congestion degree.

[0070] Of course, the current buffer data volume situation can be represented by the channel value corresponding to the buffer of the system - on - chip. The channel value corresponding to the buffer is positively correlated with the congestion degree. Among them, the larger the channel value, the less the channel is occupied and there is no congestion, indicating that the congestion condition is not reached; the smaller the channel value, the easier the channel is occupied and congestion occurs, indicating that the congestion condition is reached.

[0071] Step 112: If the current buffer data volume reaches the congestion condition, it is detected that the common link is congested.

[0072] The above method for determining whether the current buffer data volume reaches the congestion condition may include: using the current buffer data volume to determine whether the channel value corresponding to the buffer is greater than the congestion threshold, such as the threshold corresponding to the data volume; or, using the channel value corresponding to the buffer to determine whether the channel value is less than the congestion threshold, such as the threshold corresponding to the channel value. Among them, each threshold can be set by the user himself, and the congestion threshold is used to represent the congestion situation of the buffer.

[0073] Among them, the channel value of the above buffer is used to represent the buffering capacity of the buffer. The larger the channel value, the less the channel is occupied and there is no congestion.

[0074] Step 113: If the current buffer data volume does not reach the congestion condition, it is detected that the common link is not congested.

[0075] The buffer of the on-chip system in this article may include, but is not limited to, the judgment of the channel (Credit) value corresponding to the downstream buffer, and the initial value is set to the buffer size, that is, the network is unobstructed. The buffer of the on-chip system may include, but is not limited to, a buffer queue, etc. The channel corresponding to the buffer is a channel for temporary data storage and transmission.

[0076] Next, taking the channel (Credit) value corresponding to the buffer in this article as an example, first perform congestion detection to determine whether the Credit value is less than the congestion threshold. If so, congestion is detected, and it is necessary to enable the backup link and update the link status of the router, and then perform congestion detection again.

[0077] First, perform congestion detection. Determine whether Credit is less than the congestion threshold. If the Credit value is greater than the congestion threshold, it means that the link is not congested. Then determine whether the link is idle. If the link is idle, close the backup link and continue to perform congestion detection.

[0078] The common link and the backup link between two adjacent routers in this article can share a buffer. It is also possible for all the links of the on-chip system to share a buffer. As long as it can be determined that the common link between two adjacent routers is congested, no further examples will be given here.

[0079] Based on the same inventive concept as the above method, the embodiment of the present application also provides a construction device for reconstructing a network based on an on-chip system, as Figure 5 shown, the device may include the following various modules.

[0080] A status determination module 31, configured to obtain an initial network topology; the initial network topology includes: a standby link and a common link respectively connected between a current router and an adjacent next router of a pre-built on-chip system; the normal state of the standby link is a non-enabled state; An enabling detection module 32, configured to, during the process of using the common link for data transmission, in the case of detecting congestion of the common link, enable the corresponding standby link for data transmission, and update the link state of the routers in the initial network topology, to construct a current network topology including the enabled standby link and the used common link; A closing detection module 33, configured to, in the case of detecting that the congestion of the common link disappears and the link is idle, close the corresponding standby link, and update the link state of the routers in the initial network topology, to construct a current network topology including the used common link.

[0081] As an embodiment, the above device further includes: a congestion condition determination module, configured to determine that the real-time congestion situation meets the congestion decreasing condition in the following manner, which includes: the real-time congestion situation includes the change degree of the current congestion, correspondingly, the change degree of the current congestion reaches the gradually decreasing condition; or, the real-time congestion situation includes the current congestion degree; correspondingly, the current congestion degree is lower than the lowest congestion threshold.

[0082] As an embodiment, the above device further includes: a congestion detection module, configured to, during the process of using the common link for data transmission, detect whether the common link is congested in the following manner, which includes: A current buffer data volume determination unit, configured to, during the process of using the common link for data transmission, determine the current buffer data volume of the buffer of the on-chip system; A first congestion detection unit, configured to, if the current buffer data volume reaches the congestion condition, detect that the common link is congested; A second congestion detection unit, configured to, if the current buffer data volume does not reach the congestion condition, detect that the common link is not congested.

[0083] Each module / unit of the above device corresponds to the steps of the above method. The implementation processes of the functions and roles of each module / unit in the above device are specifically detailed in the implementation processes of the corresponding steps in the above method, and the same technical effects can be achieved, which will not be elaborated here.

[0084] An electronic device according to an embodiment of the present application includes the above device for constructing a reconfigured network based on an on-chip system.

[0085] The method provided by the embodiments of the present invention can be applied to an electronic device. Specifically, the electronic device can be: a desktop computer, a portable computer, a smart mobile terminal, a server, a PDA (Personal Digital Assistant), a handheld terminal, etc. Among them, the PDA can include an industrial PDA and a consumer PDA. Any electronic device that can implement the embodiments of the present invention belongs to the protection scope of the present invention and is not limited herein.

[0086] Figure 6 The following is a schematic structural diagram of an electronic device 50 provided by an embodiment of the present application.

[0087] As Figure 6 shown, the electronic device 50 includes one or more processors 51, which are used to implement the method for constructing a reconfigurable network based on a system-on-chip as described above.

[0088] In some embodiments, the electronic device 50 may include a storage medium 59. For example, a computer-readable storage medium may store a program that can be called by the processor 51, and may include a non-volatile storage medium. In some embodiments, the electronic device 50 may include a memory 58 and an interface 57. In some embodiments, the electronic device 50 may also include other hardware according to actual applications.

[0089] The computer-readable storage medium of the embodiments of the present application stores a program, and when the program is executed by the processor 51, it is used to implement the method for constructing a reconfigurable network based on a system-on-chip as described above.

[0090] The present application provides a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the method described in any one of the above is implemented.

[0091] The embodiments of the present application also provide a computer program, which is stored in a computer-readable storage medium, for example, Figure 6 the storage medium 59, and when the processor executes the computer program, it causes the processor 51 to execute the method described above.

[0092] This application may take the form of a computer program product implemented on one or more computer-readable storage media including, but not limited to, magnetic disk storage, CD-ROM, optical storage, etc., which contain program code. Computer-readable storage media include both permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information may be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0093] The above are only the preferred embodiments of this specification and are not intended to limit this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this specification shall be included within the scope of protection of this specification.

[0094] It should also be noted that the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, commodity or device. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, commodity or device including the said element.

Claims

1. A method for constructing a reconstruction network based on a system-on-chip, characterized in that, Including: Obtain an initial network topology; the initial network topology includes: a standby link and a common link respectively connected between the current router and the adjacent next router of a pre-built system-on-chip; the normal state of the standby link is a non-enabled state; During the process of data transmission using the common link, when it is detected that the common link is congested, enable the corresponding standby link for data transmission, and update the link status of the routers in the initial network topology to construct the current network topology including the enabled standby link and the used common link; When it is detected that the congestion of the common link disappears and the link is idle, close the corresponding standby link, and update the link status of the routers in the initial network topology to construct the current network topology including the used common link.

2. The construction method of the reconstruction network based on the system-on-chip according to claim 1, characterized in that During the process of data transmission using the common link, when it is detected that the common link is congested, enabling the corresponding standby link for data transmission includes: When it is detected that the common link is congested, determine the real-time congestion situation; When it is determined that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enable the corresponding standby link.

3. The method for constructing a reconfigurable network based on a system-on-chip according to claim 2, wherein, The real-time congestion situation includes the current congestion degree: When it is determined that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enabling the corresponding standby link includes: When the current congestion degree is greater than the congestion threshold, enable the standby link for data transmission.

4. The method for constructing a reconstruction network based on a system-on-chip according to claim 3, wherein, The number of the standby links is more than 1; When the current congestion degree is greater than the congestion threshold, enabling the standby link for data transmission includes: When the current congestion degree is greater than the lowest congestion threshold, enable 1 standby link for data transmission; When the current congestion degree is greater than the highest congestion threshold, enable 2 or more standby links for data transmission simultaneously or successively; the lowest congestion threshold is less than the highest congestion threshold.

5. The construction method of the reconstruction network based on the system-on-chip according to claim 2, characterized in that, The number of the standby links is more than 1; the real-time congestion situation includes the change degree of the current congestion; When it is determined that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enabling the corresponding standby link includes: When the current congestion degree is greater than the lowest congestion threshold, enable the first standby link for data transmission; When the change degree of the current congestion reaches the condition of gradually increasing, gradually enable the remaining standby links after the first standby link for data transmission.

6. The method for constructing a reconfigurable network based on a system-on-chip according to claim 2, wherein, The real-time congestion situation includes the current congestion degree: When it is determined that the real-time congestion situation meets the condition of increasing congestion, automatically and dynamically enabling the corresponding standby link includes: When it is determined that the real-time congestion situation meets the condition of increasing congestion, based on the current congestion degree, according to the corresponding relationship established between the congestion degree and the enabled number of standby links, determine the enabled number of standby links corresponding to the current congestion degree; Enable the standby links corresponding to the enabled number for data transmission.

7. The method for constructing a reconstruction network based on a system-on-chip according to claim 1, wherein The number of the standby links is more than 1; when detecting that the congestion of the common link disappears and the link is idle, closing the corresponding standby link includes: When detecting that the common link is congested, determining the real-time congestion situation; When determining that the real-time congestion situation meets the congestion reduction condition, maintaining the currently enabled standby link for data transmission until detecting that the congestion of the common link disappears and the link is idle, then closing the corresponding standby link, or gradually closing the currently enabled standby links until detecting that the congestion of the common link disappears and the link is idle, then closing all the enabled standby links.

8. The method for constructing a reconfigurable network based on a system-on-chip as claimed in claim 7, determining that the real-time congestion situation meets the congestion reduction condition in the following manner: The real-time congestion situation includes the change degree of the current congestion. Correspondingly, the change degree of the current congestion reaches the gradually decreasing condition; Or, The real-time congestion situation includes the current congestion degree. Correspondingly, the current congestion degree is lower than the lowest congestion threshold.

9. The method for constructing a reconstruction network based on a system-on-chip according to any one of claims 1 to 7, characterized in that Enabling the corresponding standby link for data transmission includes: If receiving a data packet for each data transmission, randomly selecting either the common link or the standby link from the common link and the standby link for data transmission; Or, If receiving a data packet for each data transmission, preferentially using the common link for transmission, and when the common link reaches the preset stop transmission threshold, then selecting the standby link for data transmission; Or, If receiving more than two data packets for each data transmission, separating the more than two data packets into independent sub-packets; and respectively using the standby link and the common link to transmit the independent sub-packets.

10. The method for constructing a reconfigurable network based on a system-on-chip according to any one of claims 1 to 7, characterized in that, The method further includes: During the process of using the common link for data transmission, detecting whether the common link is congested in the following manner: During the process of using the common link for data transmission, determining the current buffer data volume of the buffer of the system-on-chip; If the current buffer data volume reaches the congestion condition, detecting that the common link is congested; If the current buffer data volume does not reach the congestion condition, detecting that the common link is not congested.

11. An apparatus for constructing a reconstruction network based on a system-on-chip, characterized in that Including: A state determination module, configured to obtain an initial network topology; the initial network topology includes: standby links and common links respectively connected between the current router of the pre-constructed system-on-chip and the adjacent next router; the normal state of the standby link is the non-enabled state; An enabling detection module, configured to, during the process of using the common link for data transmission, when detecting that the common link is congested, enabling the corresponding standby link for data transmission, and updating the link state of the router of the initial network topology, and constructing a current network topology including the enabled standby link and the used common link; The closing detection module is used to close the corresponding standby link when it detects that the congestion of the common link disappears and the link is idle, and update the link status of the routers in the initial network topology, so as to construct the current network topology including the common link in the usage state.

12. An electronic device, characterized in that, It includes one or more processors for implementing the method for constructing a reconfigurable network based on a system-on-chip as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, A program is stored thereon, and when the program is executed by the processor, it implements the method for constructing a reconfigurable network based on a system-on-chip as described in any one of claims 1 to 10.

Images