Built-In Predictors for Dynamic Crosstalk Avoidance

In very deep sub-micrometer technology nodes, signal integrity of interconnects has been drastically jeopardized by crosstalk noise. To make a communication link reliable against crosstalk faults, different detection, correction, and avoidance methods have been proposed at the cost of redundant spatial and information overheads. In this paper, we propose a crosstalk prediction hardware based on an abstract model deduced from low-level interconnect evaluation for new technologies. This predictor monitors the data pattern to be sent through a communication bus and predicts those likely subjected to the crosstalk fault. Thus, to prevent crosstalk faults, the predictor dynamically engages an avoidance or detection mechanism. Top-level buses are the target of the proposed method. Simulation results reveal that the proposed communication channel is more efficient in terms of crosstalk alleviation as well as area and performance overhead compared to the state of the art reliability methods.