For automated software testing approaches like symbolic execution, path explosion is a well-known bottleneck in code analysis. Asymptotic path complexity (APC) is a metric formalizing the degree of path explosion, thus quantifying the computational difficulty of achieving path coverage. Previous work has demonstrated that APC is a more accurate and refined metric of code complexity than other common metrics such as cyclomatic or NPATH complexity. Further, APC is useful in the context of automated software testing, providing an upper bound on the growth rate of paths explored by the popular symbolic execution software Klee. Earlier algorithms computing APC only handled intraprocedural code, including non-recursive and self-recursive functions. We extend and optimize APC to measure the complexity of interprocedural programs. Our new APC algorithm, APC-IP, computes path complexity for interprocedural functions, subsumes prior approaches and is significantly more scalable.