1: \begin{abstract}

2: The quantum speed limit sets the minimum time required to transfer a quantum system completely into a given target state.

3: At shorter times the higher operation speed has to be paid with a loss of fidelity. Here we quantify the trade-off between the fidelity and the duration in a system driven by a time-varying control. The problem is addressed in the framework of Hilbert space geometry offering an intuitive 

4: interpretation of optimal control algorithms. This approach is applied to non-uniform time variations which leads to a necessary criterion for control optimality applicable as a measure of algorithm convergence. The time fidelity trade-off expressed in terms of the direct Hilbert velocity provides a robust prediction

5: of the quantum speed limit and allows to adapt the control optimization such that it yields a predefined fidelity. 

6: The results are verified numerically in a multilevel system with a constrained Hamiltonian, and

7: a classification scheme for the control sequences is proposed based on their optimizability.

8: \end{abstract}

9: