WO2020093887A1

WO2020093887A1 - Data transmission method and device for network-on-chip (noc) and electronic device

Info

Publication number: WO2020093887A1
Application number: PCT/CN2019/113429
Authority: WO
Inventors: 冯杰
Original assignee: 北京灵汐科技有限公司
Priority date: 2018-11-09
Filing date: 2019-10-25
Publication date: 2020-05-14
Also published as: CN109408257A; CN109408257B

Abstract

Disclosed are a data transmission method and device for network-on-chip (NOC) and an electronic device, used for solving the problems of high expenditure, high power consumption and low transmission efficiency when a many-core chip uses NOC for core-to-core communication. The method comprises the following steps: a first network node receives a data package comprising a destination address for representing the relative position of a target network node and a first network node in NOC (S200); in response to the determination of not being the target network node, the first network node amends the destination address in the data package according to a preset propagation direction (S201); the first network node sends the amended data package to a network node adjacent to the first network node according to the preset propagation direction, and the adjacent network node becomes a new first network node (S202).

Description

Data transmission method, device and electronic equipment for network on chip NOC

This application claims the priority of the Chinese patent application filed on November 09, 2018 with the application number 201811333210.7 and the invention titled "Data transmission method, device and electronic equipment for network-on-chip NOC", the entire content of which is cited by reference Incorporated in this application.

Technical field

The present invention relates to the field of communication technology, and in particular, to a data transmission method, device, and electronic device for network-on-chip NOC.

Background technique

With the development of artificial intelligence technology, users have higher and higher requirements on chip processing capabilities. Due to the limited processing power of single-core chips, many-core chips are more and more widely used. In the design of many-core chips, Networks on Chip (NOC) is used to realize communication between cores. Therefore, the performance of NOC is the key to the performance of many-core chips.

In the prior art, many-core chips use NOC for core-to-core communication. For example, the master data sending unit (Master, M) and multiple slave data receiving units (Slave, S) are the cores that need to communicate. When M multicasts a data packet to multiple S, the specific method is as follows: First, M adds the information of the multicast flag in the header of the data packet that it needs to send, where the multicast flag is used to indicate the To which specific destinations the data packet needs to be multicast, then M sends the data packet to the intermediate network node (Node, n) connected to it, n sends the data packet to the specific intermediate multicast directly or through other intermediate network nodes Broadcasting Station (BS), where the BS can also be called a multicast node, and after receiving the data packet, the BS searches for each in the storage area of the BS according to the information of the multicast flag in the packet header Multicast destination address, determine the number of destinations, and then the BS copies the data in the data packet according to the number of destinations, each copy of the data and the address of each multicast destination found One of the packages, Good packetizing the plurality of parallel corresponding to the multicast destination, specific, can be directly packaged data packet to the destination, may be transmitted to a destination via an intermediate network node. Specifically, as shown in Figure 1, taking one of the lines as an example, M sends the data packet to BS0 through n0 and n1, BS0 sends the packaged data packet to S0 and BS1 through n1 and n2, and BS1 passes the packaged data through n3 is sent to S1, or directly to S2.

Using the NOC in the existing technology for core-to-core communication, the data packet needs to increase the number of additional headers. As many possible multicast destinations, the number of flags needs to be added. The overhead is huge, and the BS needs Transmitting additional packet header information will bring additional loss of transmission efficiency. In many-core chips, the number of BSs is limited. If multiple Ms need to be multicast at the same time, sending data packets to S must be sent to BS first. Will increase the congestion of NOC, reduce the performance of many core chips, and will also bring additional paths, reduce transmission efficiency, and increase power consumption.

Summary of the invention

In view of this, the present invention provides a data transmission method, device and electronic device for network-on-chip NOC, which is used to solve the problem that when many core chips use NOC for core-to-core communication in the prior art, the overhead is huge 1. Large power consumption and low transmission efficiency.

According to a first aspect of the embodiments of the present invention, there is provided a data transmission method for an on-chip network NOC, including: a first network node receives a data packet, wherein the data packet includes a destination address and the destination The ground address is used to characterize the relative position of the target network node and the first network node in the NOC; in response to determining that it is not the target network node, the first network node modifies the data according to a preset propagation direction The destination address in the packet; the first network node sends the modified data packet to the network node adjacent to the first network node according to the preset propagation direction, the adjacent network The node is the new first network node.

In the embodiment of the present invention, the data packet contains only the destination address, and there is no need to add an extra number of headers in the data packet, which reduces overhead; and the first network node determines the propagation direction of the data packet according to the destination address. Using BS can achieve data multicast, improve transmission efficiency and reduce power consumption.

In one embodiment, the NOC is a one-dimensional network including multiple network nodes, and the destination address is the relative coordinates of the target network node and the first network node in the one-dimensional network.

In one embodiment, the first network node modifies the destination address in the data packet according to a preset propagation direction, which specifically includes: the first network node according to the propagation direction in the one-dimensional network, The absolute value of the coordinates in the destination address in the data packet is decreased by one.

In one embodiment, the NOC is a matrix network including multiple network nodes, and the destination address is the two-dimensional relative coordinates of the target network node and the first network node in the matrix network. The two-dimensional relative coordinates include a first direction coordinate x and a second direction coordinate y.

In one embodiment, the first network node modifies the destination address in the data packet according to a preset propagation direction, which specifically includes: the first network node responds to the first The coordinate value of the one-direction coordinate x is a positive integer greater than 0, and the propagation direction of the data packet is determined as the first direction and then the second direction according to the preset propagation direction; the first network node is based on the first In one direction, the absolute value of the coordinate x in the first direction in the destination address in the data packet is decreased by one.

In one embodiment, the first network node modifies the destination address in the data packet according to a preset propagation direction, which specifically includes: the first network node responds to the first The coordinate value of the one-direction coordinate x is a positive integer greater than 0, and the propagation direction of the data packet is determined as the second direction and then the first direction according to the preset propagation direction; the first network node is based on the first In the two directions, the absolute value of the coordinate of the second direction coordinate y in the destination address in the data packet is decreased by 1.

In one embodiment, after the first network node receives the data packet, the method further includes: in response to determining that it is the target network node, the first network node sends the data packet to the first network node. A slave data receiving unit connected to a network node.

In one embodiment, after the first network node determines that it is the target network node, the method further includes: the first network node sends the modified data packet to the site according to a preset register At least one network node specified by the register, the at least one network node being the new target network node.

In the embodiment of the present invention, the register in the network node can be dynamically configured to make the routing of the data packet very flexible. Each network node can decide whether to perform multicast according to the dynamic configuration of the register, and dynamically configure the destination of the multicast; data The packet does not need to add additional flag bits, which improves the data transmission efficiency.

According to a second aspect of the embodiments of the present invention, there is provided a data transmission device for an on-chip network NOC, including: a receiving unit for receiving a data packet, wherein the data packet includes a destination address, the The destination address is used to characterize the relative position of the target network node and the first network node in the NOC; the processing unit is configured to modify the data packet according to a preset propagation direction in response to determining that it is not the target network node The destination address; a sending unit, configured to send the modified data packet to a network node adjacent to the first network node according to the preset propagation direction, the adjacent network node is new The first network node.

In one embodiment, the processing unit is specifically configured to: reduce the absolute value of the coordinates in the destination address in the data packet by 1 according to the propagation direction in the one-dimensional network.

In one embodiment, the processing unit is specifically configured to: in response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determine the data according to the preset propagation direction The propagation direction of the packet is the first direction and then the second direction; according to the first direction, the absolute value of the coordinate x of the first direction coordinate in the destination address in the data packet is decreased by 1.

In an embodiment, the processing unit is further specifically configured to: in response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determine the The propagation direction of the data packet is the second direction before the first direction; according to the second direction, the absolute value of the coordinate of the second direction coordinate y in the destination address in the data packet is decreased by 1.

In one embodiment, after the data packet is received, the processing unit is further configured to: in response to determining that it is the target network node, the sending unit to send the data packet to the first A slave data receiving unit connected to a network node.

In one embodiment, after the response determines that it is the target network node, the sending unit is further configured to: according to a preset register, send the modified data packet to at least the register specified at least One network node, the at least one network node is a new target network node.

According to a third aspect of the embodiments of the present invention, there is provided an electronic device including: a plurality of processing cores; and an on-chip network configured to interact with data between the plurality of processing cores and external data; An instruction is stored in the plurality of processing cores, and according to the instruction, the electronic device executes the method as described in the first aspect or any possible aspect of the first aspect.

According to a fourth aspect of the embodiments of the present invention, there is provided a computer-readable storage medium on which computer program instructions are stored, and when executed by a processor, the computer program instructions implement the first aspect or any one of the first aspects Possible methods.

According to a fifth aspect of the embodiments of the present invention, there is provided a computer program instruction, which when executed on a computer, causes the computer to execute the first aspect or any possible aspect of the first aspect method.

The beneficial effects of the embodiments of the present invention include: first, the first network node receives a data packet, wherein the data packet includes a destination address, and the destination address is used to characterize the target network node and the first network node. Relative position in the NOC; in response to determining that it is not the target network node, the first network node modifies the destination address in the data packet according to a preset propagation direction; the first network node will modify The subsequent data packet is sent to a network node adjacent to the first network node according to the preset propagation direction, the adjacent network node is a new first network node, and the adjacent network node is based on The modified destination address in the received modified data packet determines that it is the target network node, and sends the modified data packet to the slave data receiving unit and other network nodes connected to it, and the other network nodes will modify After the data packet is sent to other slave data receiving unit connected to itself, if the adjacent network node determines that it is not the target network node Process again according to the above processing method. In the embodiment of the present invention, since the data packet contains only the destination address, there is no need to add an extra number of headers in the data packet, which reduces overhead; and the first network node determines the direction of the data packet according to the destination address, not BS can be used to achieve data multicast, improve transmission efficiency, and reduce power consumption.

BRIEF DESCRIPTION

The above and other objects, features and advantages of the present invention will be more clear through the following description of the embodiments of the present invention with reference to the drawings, in the drawings:

FIG. 1 is a schematic structural diagram of a data transmission provided by the prior art;

2 is a flowchart of a data transmission method for an on-chip network NOC provided by an embodiment of the present invention;

3 is a schematic structural diagram of a data transmission provided by an embodiment of the present invention;

4 is a flowchart of yet another data transmission method for an on-chip network NOC provided by an embodiment of the present invention;

5 is a schematic structural diagram of another data transmission provided by an embodiment of the present invention;

6 is a flowchart of a data transmission method for an on-chip network NOC provided by an embodiment of the present invention;

7 is a schematic diagram of a data transmission device for an on-chip network NOC provided by an embodiment of the present invention;

8 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

detailed description

The present invention is described below based on examples, but the present invention is not limited to these examples. In the following detailed description of the present invention, some specific details are described in detail. Those skilled in the art can fully understand the application without the description of these details. In addition, those of ordinary skill in the art should understand that the drawings provided herein are for illustrative purposes.

Unless the context clearly requires it, the words "including", "including" and the like in the entire specification and claims should be interpreted as the meaning of inclusive rather than exclusive or exhaustive meaning; that is, "including but not limited to" Meaning.

In the description of the present invention, it should be understood that the terms "first", "second", etc. are for descriptive purposes only, do not represent order, and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise stated, the meaning of "plurality" is two or more.

In the prior art, many-core chips use NOC for core-to-core communication. Taking the method of sending data packets from the master data sending unit to multiple slave data receiving units in the background art as an example, there are many shortcomings, as follows ： 1) Extra packet header digits need to be added to the data packet. As many multicast destinations are possible, as many flag bits need to be added. For example, any core in a multi-core chip composed of a thousand cores may be Multicast the data packet to other cores, then the packet header will increase the flag bit of all other cores except itself, that is 1000-1 = 999 bits, the overhead is huge, and the transmission of additional packet header information will bring additional efficiency transmission loss ; 2) Since the multicast node BS is used to send the data packet, a special storage area for storing all possible destination addresses of the multicast node needs to be set in the BS, for example, consisting of a thousand cores Among the many-core chips, if any multicast node may need to multicast the received data packet to one hundred destination cores, the address of the one hundred destination cores needs to be saved, resulting in the storage of the BS The overhead is huge. In the entire many-core chip, there may be many of the above-mentioned multicast nodes, which leads to a greater storage overhead for the entire many-core chip; 3) The BS uses a dedicated circuit, and it cannot reuse the circuit with other network nodes. Circuit area; 4) The BS inquires the destination address of the multicast according to the information of the multicast flag in the header of the received data packet, and then repackages the destination address with the valid data in the data packet and sends it out. Added additional query and packaging, increased power consumption; 5) In a many-core chip, the number of BS is limited, if there are multiple M need to be multicast at the same time, it will increase the NOC congestion and reduce the entire crowd The performance of the core chip; 6) When sending data packets to certain destinations, the use of BS multicast will bring additional paths, reduce the efficiency of routing, increase the time of data packet delivery, and increase the consumption of data transmission Power consumption.

In order to solve the problems of the prior art, the present invention provides a data transmission method for network-on-chip NOC, as shown in FIG. 2, specifically including:

Step S200: The first network node receives a data packet, where the data packet includes a destination address, and the destination address is used to characterize the relative position of the target network node and the first network node in the NOC.

Specifically, the data packet may be sent by the main data sending unit, or may be sent by another network node. The above data packet is packaged according to a conventional data packet format, and the data packet header contains the ordinary destination address and other Information, there is no need to increase the number of additional headers.

Step S201: In response to determining that it is not the target network node, the first network node modifies the destination address in the data packet according to a preset propagation direction.

Specifically, the preset propagation direction is divided into two cases.

Case 1: The NOC is a one-dimensional network including multiple network nodes, and the destination address is the relative coordinates of the target network node and the first network node in the one-dimensional network. For example, suppose The destination address is (a), where the absolute value of a is greater than 0.

Case 2: The NOC is a matrix network including multiple network nodes, and the destination address is a two-dimensional relative coordinate of the target network node and the first network node in the matrix network. The relative coordinates include a first direction coordinate x and a second direction coordinate y. The propagation direction of the data packet is the first direction and then the second direction, or the second direction and then the first direction.

For example, the destination address is (x, y), where the absolute values of x and y are greater than 0, respectively. The absolute values of x and y represent the distance traveled, and the symbols represent the direction of propagation. The transmission process is the process of decreasing the absolute value of x and y. X and y can be arbitrarily set according to actual relative positions, which is not limited in the embodiment of the present invention.

In the embodiment of the present invention, a specific example is given for the first network node to modify the destination address in the data packet according to the preset propagation direction, which is as follows: For the first case above, assume the value of a Is 3, that is, the destination address in the data packet is (3), indicating that the first network node needs to pass through two intermediate network nodes to reach the target network node (0), because the situation one is fixed unidirectional propagation in a one-dimensional network Therefore, the first network node modifies the destination address in the data packet to (2).

For the second case and the first case, suppose that the propagation direction is determined to be the first direction and then the second direction, the value of x is 3, the value of Y is 2, and the destination address in the data packet is (3, 2), where , (3, 2) is the relative position of the first network node and the target network node (0, 0). Since the propagation direction is the first direction and then the second direction, the first network node converts the destination in the data packet. The address is modified to (2, 2); in the second case, it is assumed that the propagation direction is determined to be the second direction before the first direction. Therefore, the first network node modifies the destination address in the data packet to (3, 1). In the embodiment of the present invention, when the absolute values of x and y are respectively greater than 0, the propagation direction can be arbitrarily set, which is not limited in the embodiment of the present invention.

Step S202: The first network node sends the modified data packet to a network node adjacent to the first network node according to the preset propagation direction, and the adjacent network node is a new first Network node.

For example, for the first case above, the modified data packet is sent to a network node adjacent to the first network node according to a preset propagation direction, and the relative position of the adjacent network node and the target network node is (2) . For the second case and the first case, the modified data packet is sent to the adjacent network node with the destination address (2, 2) of the first network node, the adjacent network node and the target network node The relative position is (2, 2). For the second and second cases above, send the modified data packet to the adjacent network node with the destination address (3, 1) of the first network node, the adjacent network node and the target network node The relative position is (3, 1), and since the destination address is not 0, it means that the adjacent network node is not the target network node, and the adjacent network node serves as a new first network node and returns to S200 for retransmission.

In the embodiment of the present invention, first, the first network node receives a data packet, wherein the data packet includes a destination address, and the destination address is used to characterize the target network node and the first network node in the NOC Relative position; in response to determining that it is not the target network node, the first network node modifies the destination address in the data packet according to a preset propagation direction, the first network node translates the modified data packet according to The preset propagation direction is sent to a network node adjacent to the first network node. The adjacent network node is a new first network node. Since the data packet contains only the destination address, the data packet There is no need to increase the number of additional packet headers, reducing overhead; and the first network node determines the direction of data packet propagation based on the destination address, without using BS to achieve data multicast, improve transmission efficiency, reduce power consumption .

Optionally, after step S200, the method further includes: the first network node determining whether it is the target network node according to the destination address in the data packet.

Optionally, after step S200, the method further includes: in response to determining that it is the target network node, the first network node sends the data packet to a slave data receiving unit connected thereto.

Specifically, assuming that the propagation direction of the data packet is fixed unidirectional propagation in a one-dimensional network, and the destination address in the received data packet is (0), it is determined that it is the target network node; assuming the propagation direction matrix of the data packet In the network, the first direction is followed by the second direction or the second direction is followed by the first direction. If the destination address in the received data packet is (0, 0), it is determined that it is the target network node.

Optionally, while the first network node sends the data packet to the slave data transfer unit connected thereto, it also includes sending the data packet to at least one network node designated by the register according to a preset register, Among them, the register can also set whether to change the destination address, if it does not change, the next network node can send the data packet to the slave data receiving unit connected to it after receiving the data packet, and can also send the data packet according to the setting of the register Send to other network nodes; if the register setting changes the destination address, return to step S200 to continue processing.

For example, suppose that the register in the first network node is 3 digits, corresponding to the three network nodes adjacent to it. When the value of the register is the set value, the received data packet is sent to the pair without modification For a network node used, the set value may be 1 or 0, which is not limited in the embodiment of the present invention. For example, if the setting value is 1, it means that it needs to be sent to the corresponding network node. If the setting value is 0, it means that it does not need to be sent to the corresponding network node. When the value of the register in the first network node is 110, it corresponds to the middle Nodes A, B, and C indicate that the data packets are sent to A and B without modification, and the data packets are not sent to C. After receiving the data packets, A and B serve as new target network nodes, and perform corresponding actions according to their pre-set registers. Operation.

The following uses two specific embodiments to describe in detail a data transmission method for an on-chip network NOC provided by an embodiment of the present invention.

Specific example one

Assuming that the propagation direction of the data packet is a fixed one-way propagation, a schematic diagram of the specific sending process of the data packet. As shown in FIG. 3, the master data sending unit M multicasts the data packet to the slave data receiving units S0, S1 and S2, n0 receives To the data packet sent by M, send the data packet to n1, n1 send the data packet to S0 and n3, n3 send the data packet to S1 and n4, n4 send the data packet to S2, the specific processing flow As shown in Figure 4:

Step S400: The network node n0 receives the first data packet sent by M, and determines that the destination address in the first data packet is (1).

Step S401: Determine that the preset propagation direction of the network node n0 is fixed one-way propagation.

Step S402: The network node n0 modifies the destination address in the first data packet to (0), and determines the data packet with the modified address (0) as the second data packet.

Step S403: The network node n0 sends the second data packet to the network node n1.

Step S404: The network node n1 receives the second data packet.

Step S405: The network node n1 determines that it is the target network node according to the destination address (0) in the second data packet.

Step S406: The network node n1 sends the second data packet to the connected S0, and sends the second data packet to the network node n3 without modification according to the setting of its own register.

Step S407: The network node n3 determines that it is the target network node according to the destination address (0) in the second data packet.

Step S408: The network node n3 sends the second data packet to S1 connected thereto, and sends the second data packet to the network node n4 without modification according to the setting of its own register.

Wherein, due to the setting of the register, the second data packet is not sent to the network node n2.

Step S409: The network node n4 determines that it is the target network node according to the destination address (0) in the second data packet.

Step S410: The network node n4 sends the second data packet to S2 connected thereto.

Specific embodiment two

Assuming that the core C0 multicasts data to the cores C4, C5 and C7, the specific schematic diagram is shown in FIG. 5, wherein the multicast data may be the processing unit circuit or the main data sending unit M, and the propagation direction is the first X axis After the Y axis, the specific processing flow is shown in Figure 6.

Step S600: The network node n0 receives the first data packet sent by the core C0, and determines that the destination address in the first data packet is (1, 1).

Step S601: Determine that it is not the target network node according to the destination address, and determine that the preset propagation direction of the network node n0 is the first X axis.

Step S602: The network node n0 modifies the destination address in the first data packet to (0, 1), and determines the modified data packet with the address (0, 1) as the second data packet.

Step S603: The network node n0 sends the second data packet to the network node n1.

Step S604: The network node n1 receives the second data packet.

Step S605: The network node n1 determines that it is not the target network node according to the destination address (0, 1) in the second data packet.

Step S606: The network node n1 modifies the destination address in the second data packet to (0, 0), and determines the modified data packet with the address (0, 0) as the third data packet.

Specifically, after receiving the second data packet, the network node n1 first adopts the strategy of transmitting in the X-axis direction, but at this time, the X of the destination address in the packet header is already 0, so the transmission direction is changed to the Y-axis.

Step S607: The network node n4 receives the third data packet.

Step S608: The network node n4 determines that it is the target network node according to the destination address (0, 0) in the third data packet, sends the third data packet to the core C4 connected to it, and sends the third data packet according to the setting of its own register Three data packets are sent to network node n5 and network node n7.

Specifically, the network node n4 determines that the value of the bit corresponding to the network node n5 and the network node n7 in its register is 1, indicating that the third data packet needs to be sent to the network node n5 and the network node n7 without any modification. The value of the bit corresponding to the network node n3 in the register is 0, so no data packet is sent to n3.

Step S609: The network node n5 receives the third data packet.

Step S610: The network node n5 determines that it is the target network node according to the destination address (0, 0) in the third data packet, and sends the third data packet to the core C5 connected to it; the network node n5 according to the setting of its own register It is determined that the values of the adjacent network nodes are all 0, and then the transmission is stopped.

Step S611: The network node n7 receives the third data packet.

Step S612, the network node n7 determines that it is the target network node according to the destination address (0, 0) in the third data packet, and sends the third data packet to the core C7 connected to it; the network node n7 according to the setting of its own register Judging that the value of the bit of the adjacent network node n6 is 1, the network node n6 repeats this step to continue sending.

Among them, step S609 and step S611 are performed simultaneously.

The data transmission method for the on-chip network NOC proposed by the embodiment of the present invention also has the following advantages. The control circuit or core of the many-core chip can dynamically configure the registers in the network node at any time, making the routing of data packets very flexible. Network nodes can decide whether to perform multicast and dynamically configure the destination of multicast according to the dynamic configuration of registers; therefore, data packets do not need to add additional flag bits, which improves data transmission efficiency; intermediate network nodes do not require any additional storage addresses Space reduces the chip area; no additional unpacking and packaging is required during multicast, which reduces the delay of data transmission; data packets do not add additional routing paths, reducing the delay of data transmission; data packets do not Certain nodes must be passed to reduce the congestion of many-core chips.

7 is a schematic diagram of a data transmission device for an on-chip network NOC provided by an embodiment of the present invention. As shown in FIG. 7, the data transmission device for the on-chip network NOC of this embodiment includes: a receiving unit 71, a processing unit 72, and a sending unit 73, wherein the receiving unit 71 is used to receive a data packet, wherein the data The packet contains a destination address, which is used to characterize the relative position of the target network node and the first network node in the NOC; the processing unit 72 is used to respond to determining that it is not the target network node, according to The set propagation direction modifies the destination address in the data packet; the sending unit 73 is configured to send the modified data packet to a network adjacent to the first network node according to the preset propagation direction Node, the adjacent network node is a new first network node.

Optionally, the NOC is a one-dimensional network including multiple network nodes, and the destination address is the relative coordinates of the target network node and the first network node in the one-dimensional network.

Optionally, the processing unit is specifically configured to: reduce the absolute value of the coordinates in the destination address in the data packet by 1 according to the propagation direction in the one-dimensional network.

Optionally, the NOC is a matrix network including multiple network nodes, and the destination address is a two-dimensional relative coordinate of the target network node and the first network node in the matrix network. The dimensional relative coordinates include a first direction coordinate x and a second direction coordinate y.

Optionally, the processing unit is specifically configured to: in response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determine the data packet's The propagation direction is the first direction before the second direction; according to the first direction, the absolute value of the coordinate x of the first direction coordinate in the destination address in the data packet is decreased by 1.

Optionally, the processing unit is further specifically configured to: in response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determine the data packet according to the preset propagation direction The propagation direction of is the second direction before the first direction; according to the second direction, the absolute value of the coordinate of the second direction coordinate y in the destination address in the data packet is decreased by 1.

Optionally, after receiving the data packet, the processing unit is further configured to: in response to determining that it is the target network node, the sending unit to send the data packet to the first network node Connected slave data receiving unit.

Optionally, after the response determines that it is the target network node, the sending unit is further configured to send the modified data packet to at least one network specified by the register according to a preset register Node, the at least one network node is a new target network node.

8 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. As shown in FIG. 8, the electronic device of this embodiment includes a processing core 11-1N and a network on chip 14. The processing cores 11-1N are all connected to the on-chip network 14. The on-chip network 14 is used to exchange data between the N processing cores and external data. The N processing cores store instructions, and according to the instructions, the electronic device performs the following operations: a first network node receives a data packet, where the data packet includes a destination address, and the destination address is used for Characterize the relative position of the target network node and the first network node in the NOC; in response to determining that it is not the target network node, the first network node modifies the data in the data packet according to a preset propagation direction Destination address; the first network node sends the modified data packet to the network node adjacent to the first network node according to the preset propagation direction, the adjacent network node is a new A network node.

As those skilled in the art will appreciate, various aspects of embodiments of the present invention may be implemented as a system, method, or computer program product. Therefore, various aspects of the embodiments of the present invention may take the form of a complete hardware implementation mode, a complete software implementation mode (including firmware, resident software, microcode, etc.) or generally referred to as "circuits" or "modules" herein. "Or" system "that combines software aspects with hardware aspects. In addition, various aspects of embodiments of the present invention may take the form of a computer program product implemented in one or more computer-readable media, the computer-readable medium having computer-readable program code implemented thereon.

Any combination of one or more computer-readable media may be utilized. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing. A more specific example (not an exhaustive list) of computer-readable storage media will include the following: electrical connection with one or more wires, floppy disk for portable computer, hard disk, random access memory (RAM), read-only memory ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the foregoing. In the context of an embodiment of the present invention, a computer-readable storage medium may be any tangible medium that can contain or store a program used by or in conjunction with an instruction execution system, device, or device.

The computer-readable signal medium may include a propagated data signal having computer-readable program code implemented therein as in baseband or as part of a carrier wave. Such a propagated signal can take any of a variety of forms, including but not limited to: electromagnetic, optical, or any suitable combination thereof. The computer-readable signal medium may be any computer-readable medium that is not a computer-readable storage medium and that can communicate and propagate programs used by or in conjunction with an instruction execution system, device, or device Or transmission.

The program code implemented on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for performing operations directed to various aspects of embodiments of the present invention may be written in any combination of one or more programming languages, including: object-oriented programming languages such as Java, Smalltalk, C ++, etc .; And conventional process programming languages such as the "C" programming language or similar programming languages. The program code may be executed entirely on the user's computer, partly on the user's computer as an independent software package; partly on the user's computer and partly on a remote computer; or entirely on the remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using the Internet service provider's Internet) .

The above-mentioned flowchart illustrations and / or block diagrams of the method, device (system) and computer program product according to the embodiments of the present invention describe various aspects of the embodiments of the present invention. It will be understood that each block of the flowchart illustration and / or block diagram and a combination of blocks in the flowchart illustration and / or block diagram can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general-purpose computer, special-purpose computer, or other programmable data processing device to produce a machine so that instructions (executed via the processor of the computer or other programmable data processing device) are created for implementation Flowchart and / or block diagram block or means of function / action specified in the block.

These computer program instructions can also be stored in a computer-readable medium that can instruct a computer, other programmable data processing device, or other apparatus to operate in a specific manner, so that the generation of instructions stored in the computer-readable medium includes implementation in flowcharts and / Or block diagrams or artifacts of specified functions / actions in the blocks.

Computer program instructions can also be loaded onto a computer, other programmable data processing device, or other device to cause a series of operable steps to be performed on the computer, other programmable device, or other device to produce a computer-implemented process so that the computer The instructions executed on other programmable devices provide procedures for implementing the functions / acts specified in the flowchart and / or block diagram blocks or blocks.

The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

A data transmission method for network-on-chip NOC, characterized in that it includes:

The first network node receives a data packet, where the data packet includes a destination address, where the destination address is used to characterize the relative position of the target network node and the first network node in the NOC;

In response to determining that it is not the target network node, the first network node modifies the destination address in the data packet according to a preset propagation direction;

The first network node sends the modified data packet to a network node adjacent to the first network node according to the preset propagation direction, and the adjacent network node is a new first network node.
The method according to claim 1, wherein the NOC is a one-dimensional network including a plurality of network nodes, and the destination address is the target network node and the first network node in the one-dimensional Relative coordinates in the network.
The method according to claim 2, wherein the first network node modifies the destination address in the data packet according to a preset propagation direction, specifically including:

The first network node decrements the absolute value of the coordinates in the destination address in the data packet according to the direction of propagation in the one-dimensional network.
The method of claim 1, wherein the NOC is a matrix network including a plurality of network nodes, and the destination address is the target network node and the first network node in the matrix network The two-dimensional relative coordinates of, the two-dimensional relative coordinates include a first direction coordinate x and a second direction coordinate y.
The method according to claim 4, wherein the first network node modifies the destination address in the data packet according to a preset propagation direction, specifically including:

In response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, the first network node determines that the propagation direction of the data packet is first according to the preset propagation direction The second direction after the direction;

According to the first direction, the first network node decrements the absolute value of the coordinate x of the first direction coordinate x in the destination address in the data packet.
The method according to claim 4, wherein the first network node modifies the destination address in the data packet according to a preset propagation direction, specifically including:

In response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, the first network node determines that the propagation direction of the data packet is first according to the preset propagation direction The first direction after the direction;

The first network node decrements the absolute value of the coordinate of the second direction coordinate y in the destination address in the data packet according to the second direction.
The method according to claim 1, wherein after the first network node receives the data packet, the method further comprises:

In response to determining that it is the target network node, the first network node sends the data packet to the slave data receiving unit connected to the first network node.
The method according to claim 7, wherein after the first network node determines that it is the target network node, the method further comprises:

The first network node sends the modified data packet to at least one network node specified by the register according to a preset register, and the at least one network node is the new target network node.
A data transmission device for network-on-chip NOC, characterized in that it includes:

A receiving unit, configured to receive a data packet, wherein the data packet includes a destination address, and the destination address is used to characterize the relative position of the target network node and the first network node in the NOC;

A processing unit, configured to modify the destination address in the data packet according to a preset propagation direction in response to determining that it is not the target network node;

The sending unit is configured to send the modified data packet to a network node adjacent to the first network node according to the preset propagation direction, and the adjacent network node is a new first network node.
The apparatus according to claim 9, wherein the NOC is a one-dimensional network including a plurality of network nodes, and the destination address is the target network node and the first network node in the one-dimensional Relative coordinates in the network.
The apparatus of claim 10, wherein the processing unit is specifically configured to:

According to the propagation direction in the one-dimensional network, the absolute value of the coordinates in the destination address in the data packet is decreased by one.
The apparatus according to claim 9, wherein the NOC is a matrix network including a plurality of network nodes, and the destination address is the target network node and the first network node in the matrix network The two-dimensional relative coordinates of, the two-dimensional relative coordinates include a first direction coordinate x and a second direction coordinate y.
The apparatus of claim 12, wherein the processing unit is specifically configured to:

In response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determining the propagation direction of the data packet as the first direction and then the second direction according to the preset propagation direction;

According to the first direction, the absolute value of the coordinate x of the first direction in the destination address in the data packet is decreased by one.
The apparatus according to claim 12, wherein the processing unit is further configured to:

In response to the coordinate value of the first direction coordinate x in the destination address being a positive integer greater than 0, determine that the propagation direction of the data packet is the second direction before the first direction according to the preset propagation direction;

According to the second direction, the absolute value of the coordinate of the second direction coordinate y in the destination address in the data packet is decreased by 1.
The apparatus according to claim 9, wherein after the data packet is received, the processing unit is further configured to:

In response to determining that it is the target network node, the sending unit is configured to send the data packet to a slave data receiving unit connected to the first network node.
The apparatus of claim 15, wherein after the response determines that it is the target network node, the sending unit is further configured to:

Sending the modified data packet to at least one network node designated by the register according to a preset register, where the at least one network node is a new target network node.
An electronic device, characterized in that the electronic device includes:

Multiple processing cores; and

The on-chip network is configured to interact with the data between the multiple processing cores and external data;

An instruction is stored in the plurality of processing cores, and the electronic device executes the method according to any one of claims 1-8 according to the instruction.
A computer-readable storage medium on which computer program instructions are stored, characterized in that, when executed by a processor, the computer program instructions implement the method according to any one of claims 1-8.
A computer program instruction, characterized in that, when the computer program instruction runs on a computer, it causes the computer to execute the method according to any one of claims 1-8.