Argonne National Laboratory (ANL) Argonne, United States of America
In this work, we present a new large-scale parallel quantum circuit simulator. It is based on the tensor network contraction technique to represent quantum circuits. We propose a novel late parallelization algorithm to split tensor expressions. By using our MPI/OpenMP parallelization scheme, we have significantly increased the size of circuits we can compute, as well as the speed of calculations.
The simulator benchmarking studies were performed for Quantum Approximate Optimization Algorithm (QAOA) circuits. We computed 210 qubit QAOA circuits with depth 1,785 gates on 1,024 nodes and 213 TB of memory on the Theta supercomputer, which is based on Intel Xeon Phi architecture. To the best of our knowledge, this constitutes the largest QAOA quantum circuit simulations reported to this date.