US Patent 11983515 Optimizing source code using callable unit matching

An embodiment searches source for callable units, and scans the contents of the callable units, and generates a set of unit groups that each include one or more of the callable units. The embodiment preprocesses the callable units and then generates hashes of each of the callable units. The embodiment generates a data structure comprising unit identifiers associated with the hashes and their corresponding callable units. The embodiment generates statistical data using a callable-unit metric and updates the data structure by prioritizing the callable units based on the statistical data. The embodiment compares hashes of the callable units and identifies first and second callable units as redundant callable units. The embodiment updates the data structure by identifying the second callable unit as being redundant to the first callable unit. The embodiment generates a report of redundant callable units based on the data structure.