The ITensor Book
Tensor methods are a powerful approach for problems in physics and applied math, but keeping track of tensor indices by hand can be tedious and make your code fragile and prone to bugs.
Inspired by diagrammatic notation for tensor networks, ITensor lets you focus on the connectivity of tensor networks without thinking about low-level details like index ordering and data permutation. Indices in ITensor carry extra information so that "multiplying" two ITensors contracts all matching indices, similar to Einstein summation. ITensor is very efficient despite these high level features.
The goal of this book is to quickly familiarize you with the main features of ITensor. Later chapters turn to advanced features and customizing higher-level parts of the library.
Table of Contents
Introduction to ITensor
QN ITensors: Block-Sparse, Quantum Number Conserving Tensors
Matrix Product States and DMRG
Design of ITensor Library and Internals
Back to Main