A fault-tolerant adaptive routing method based on ant colony algorithm

Abstract

The invention discloses a fault-tolerant self-adaptive routing method based on ant colony algorithm, which includes: marking the position of each node in a two-dimensional network and setting the routing table; detecting a faulty node, if there is a faulty node, calculating the number of nodes adjacent to the faulty node The fault value of the node and the effective buffer length; if there is no faulty node, use the adaptive routing algorithm to find and obtain one or more routing paths to be selected; select an optimal routing path from the routing paths to be selected according to the evaluation scale, Update the pheromone routing table according to the ant colony routing algorithm; check whether the destination node is reached, and if so, end the routing; otherwise, judge whether there is a deadlock in the routing path, and end if there is a deadlock; otherwise, re-detect the faulty node. Based on the ant colony algorithm, the path selection mechanism in the fault-tolerant adaptive routing algorithm is optimized and the pheromone routing table is updated. In the case of faulty nodes, the data transmission delay of the entire network on chip is reduced, and the network load is balanced.

Description

technical field [0001] The invention relates to the technical field of routing of on-chip networks, in particular to a fault-tolerant adaptive routing method based on an ant colony algorithm. Background technique [0002] With the development of semiconductor technology, chips are gradually developing from single-core to multi-core and even many-core, which requires more and more attention to the interconnection of the system when designing the chip architecture. However, the current mainstream bus-based interconnection method can no longer meet the needs of the growing chip scale. In response to this situation, researchers have proposed a chip interconnection method based on a network on chip. In an on-chip network, if the network scale is small, deterministic routing is generally used because it is simple to implement and can meet performance requirements; as the network scale increases, adaptive routing algorithms are used because deterministic routing has a large It may...

AI Technical Summary

Problems solved by technology

The fault-tolerant adaptive routing algorithm solves this problem from the routing algorithm level. Traditional fault-tolerant adaptive algorithms generally use a single value for fault reminders. At the same time, the storage overhead of the routing table is too large, and a large amount of time is spent on routing path selection. time overhead

Although the effect of reducing the delay is achieved to a certain extent, there are also certain limitations.

Images