Narx architecture for task contention models using time-series and fine-grain instrumentation for mpsocs

Abstract

A computer-implemented method of predicting a timeseries of time delays resulting from contention between tasks running in parallel on a multi-processor system using a trained Machine Learning based Task Contention Model may include executing a plurality of actual execution tasks on the multi-processor system in isolation, and for each task and during execution of the respective task, capturing a timeseries including a plurality of uWindows. The method may also includes inferring, from a time-agnostic ML based regressor, a predicted contention delay for the first uWindow of the timeseries when two or more of the plurality of actual execution tasks are executed on parallel on the multi-processor system. The method further includes inferring, from a Non-linear AutoRegressor with exogenous inputs (NARX), predicted contention delays for each subsequent uWindow of the timeseries based on the respective captured snapshot, and the predicted values for the previous time periods fed back into the NARX.