Hmmm.... I think it depends on the context. For example, something I've been thinking about recently for my research: in K-Kbar mixing in the Standard Model, you can have effects from heavy quarks and light quarks running through loops. The heavy quarks are highly virtual (such as the top and charm) and so, thanks to the asymptotic freedom of QCD, these contributions are perturbative. However, the contribution from the light quark masses (up) is definitely not perturbative, since these quarks are quite light and are not "as virtual" as the heavy quarks. This means that they can exist longer from the uncertainty principle, and this means that the contributions from light quarks is ACTUALLY a contribution from pions, the IR limit of QCD. These are called "long-distance" contributions, and cannot be computed with perturbative QCD.