The most optimal mode of supercomputer utilization is continuous launching of tasks occupying the entire counting field. However, in reality, such a mode is practically uncommon, except for cases when the supercomputer is actually built to solve a single task. This is due to various reasons: limitations on the scalability of tasks, conditions on the number of processes (e.g., power of two), etc. Even if all the tasks can occupy the entire counting field, it is necessary to start a new task immediately after the previous one is finished. It is not easy to track this manually.