1: \begin{abstract}

2: Given in the title are two algorithms to compute the extreme eigenstate of a high-dimensional Hermitian matrix using the tensor train (TT) / matrix product states (MPS) representation.

3: Both methods empower the traditional alternating direction scheme with the auxiliary (e.g. gradient) information, which substantially improves the convergence in many difficult cases.

4: Being conceptually close, these methods have different derivation, implementation, theoretical and practical properties.

5: We emphasize the differences, and reproduce the numerical example to compare the performance of two algorithms.

6: \par {\it Keywords:} high--dimensional problems, DMRG, MPS, tensor train format, extreme eigenstate.

7: \par {\it MSC:}

8: 15A18,  % Eigenvalues and eigenvectors

9: 15A69, %  Multilinear algebra, tensor products

10: 65F10, %  Iterative methods for linear systems [See also 65N22]

11: 65F15, %  Eigenvalues, eigenvectors

12: 82B28, %  Renormalization group methods

13: 82B20      %  Lattice systems (Ising, dimer, Potts, etc.) and systems on graphs

14: \par {\it PACS:}

15: 02.10.Xm,   % Multilinear algebra

16: 02.60.Dc,   % Numerical linear algebra

17: 75.10.Pq,   % Spin chain models

18: 05.10.Cc    % Renormalization group methods

19: \end{abstract}

20: