In this presentation, I will describe our group’s efforts in leveraging high-performance computing (HPC), particularly the RMACC cluster, to tackle computationally intensive optimization problems in quantum computing. In our first case study, we address a quantum optimal control problem that requires fine-tuning numerous pulse parameters to optimize the performance of a quantum gate on a superconducting quantum computer. This optimization is carried out through large-scale parallel executions of a stochastic gradient descent (SGD) algorithm with multiple random seeds on RMACC. In the second case, we apply RMACC to solve a maximum likelihood estimation (MLE) problem for learning structured quantum states from synthetic measurement data. Our approach involves generating a vast number of measurement samples in parallel using a novel autoregressive method and subsequently performing MLE via SGD. In both applications, our approach yields near-optimal results that align with theoretical upper bounds, demonstrating that RMACC could provide an efficient, cost-effective HPC solution for state-of-the art quantum research in local institutions.
