Efficient Classical Simulation of the Quantum Random Access Memory (2503.13832v1)

Abstract: Quantum Random Access Memory (QRAM), despite its fundamental role in quantum information processing, has yet to be experimentally realized. Given this challenge, classical simulations serve as essential tools for gaining deeper insights into the underlying physical mechanisms of quantum systems and for advancing the development of scalable, and reliable quantum devices. However, general-purpose simulation methods become impractical due to exponential memory growth and rising computational costs, leaving an open gap for tailored approaches to simulate QRAM efficiently. Here, we present an efficient and scalable framework for simulating both noiseless and noisy bucket brigade QRAM circuits. Our approach leverages a branch-wise encoding scheme, structuring simulations around query branches rather than full circuits. This encoding enables layer-wise operations and pruning algorithms, significantly improving memory efficiency and simulation speed, particularly in noisy scenarios. As a result, our framework achieves linear computational scaling in the noiseless case and maintains stable performance under realistic noise conditions. On a single workstation, we successfully simulate a 20-layer full-address QRAM in under two hours using less than 1 GB of memory. Additionally, we extend noisy QRAM simulations to 30 layers with $2^{10}$ branches, surpassing the scale of previously reported QRAM simulations. Furthermore, we demonstrate the integration of our simulator with currently existing quantum simulators by incorporating error filtration techniques to explore noise suppression strategies in QRAM architectures. These results critically assess QRAM's feasibility, highlighting both its potential and limitations in noise suppression, while positioning our simulator as a key tool for evaluating future QRAM-based algorithms and error suppression strategies.