Graph-based Projected Entangled-Pair State with GPU Acceleration

Quantum algorithms promise to provide speedups or even solutions to previously unfeasible problems. As quantum computing matures, benchmarks are needed as a standardized method to evaluate and compare different quantum computing platforms. Hamiltonian evolution serves as a well studied problem setting for this purpose. To verify quantum solutions, classical simulations of benchmarking computations are desired. However, simulating circuits with many qubits on classical hardware can be computationally expensive.
Popular frameworks like Qiskit, Cirq and Tensorflow Quantum use state vector simulation, which can handle systems with high entanglement but are limited by the number of qubits. Tensor networks, however, can efficiently represent large quantum states, as long as the amount and structure of entanglement is limited. This is a well suited setting for approximating ground states of Hamiltonians, which are slightly entangled quantum states. One particular approach is the graph-based Projected Entangled-Pair State (gPEPS) method, which uses the graph structure of the underlying hardware topology. GPUs, with their parallel computing power, have proven valuable in matrix and tensor calculation and will be useful in speeding up the simulation of quantum circuits. 
The goal of this master thesis is to investigate the possibilities of combining the tensor network method gPEPS with the parallel computing power of GPUs. This will be implemented using NVIDIA's CuTensorNet.