Network path selection method for heat balance sheet with fault tolerance function

Abstract

The invention discloses a network path selection method for heat balance sheet with fault tolerance function, which mainly solves the problem that the prior art does not have fault tolerance function and can not guarantee the packet delay property relying on the routing table. The execution steps are as follows: (1) reading the current ode address information and judging the current node type; (2) reading the packet target node address information and judging the packet priority; (3) determining the routing output port set of the shortest path according to the node address information, the node type and the packet priority; (4) defining the selection cost function of output port: i=1, 2, reading the calculation parameter of cost function and calculating the selection cost of output port; and (5) determining the output port. The invention has the advantages of improving the network-on-chip heat balance, reliability and service life and effectively satisfying the requirement of network-on-chip property to the practical application.

Description

technical field [0001] The invention belongs to the technical field of network communication, and relates to NoC communication of a network on chip, in particular to a method for selecting a path of a heat balance network on chip with a fault tolerance function. The method can realize high-efficiency data transmission of the on-chip network, ensure the heat balance of the on-chip network, improve the service life of the chip, and simultaneously meet various requirements of fault tolerance, load balancing and service quality. Background technique [0002] With the rapid development of semiconductor integrated circuit technology and the continuous progress of new semiconductor material technology, SoC has become the focus of the microelectronics field. However, with the sharp increase of chip integration, the system-on-chip has become more and more complex, and has been greatly challenged in terms of delay, throughput, energy consumption, reliability, and scalability. The pro...

AI Technical Summary

Problems solved by technology

The shortcomings of this method are: first, source routing cannot adapt to changes in network status, and changes in traffic will cause network performance to degrade; second, the area and energy consumption of the network on chip are strictly limited, and the routing table will increase the area and energy consumption. The consumption overhead is not suitable for on-chip networks; third, load balancing during the entire working time is not equivalent to temperature balancing, and processing a large number of packets in a short period of time will also cause the temperature of the node to be too high; fourth, it does not have fault tolerance. The entire chip does not work after a failure

The shortcomings of this method are: first, closing the entire node layer has a certain blindness; second, the packet delay on the closed node layer is relatively large; third, the temperature detector increases the area overhead and the router structure. the complexity

Images