The ITensor library supports the productive development of robust and efficient software based on tensor networks. The innovative design of the ITensor library lets users focus on the connectivity of tensor networks instead of lower-level considerations, and is modeled on tensor diagram notation. ITensor has a multi-layered design including basic dense tensors; sparse tensor types; sophisticated handling of quantum number symmetries; and high-level algorithms for matrix product states.
In addition to providing ongoing support for ITensor and its users, there is a major effort at CCQ to continue expanding the scope of ITensor to make handling complex quantum models easier and to continue incorporating the latest algorithmic developments and tensor network formats.
Project Leader: Miles Stoudenmire