Inspired by the insight of quantum entanglement in the perspective of quantum information theory, a class of wave functions, namely tensor network states (TNS) are proposed to describe the strongly correlated systems. Especially in 2D, the projected entangled pair states (PEPS) and multi-entanglement renormalization ansatz (MERA) have been proved to be powerful simulation methods to explore the strongly correlated systems.

However, due to the high computation cost and complexity in coding of TNS algorithm, an easy-to-use tensor network package with high efficienty is still in demand. The Tensor Network State Package (TNSP) is developed to this end. It is built on mature math library(MKL/lapack). While retaining the efficienty of the low level math library, it provides a highly user-friendly interface and a robust error-detecting mechanism. Besides, tensors with abelian symmetry as well as with fermionic commutating rule are fully supported.