Hybrid Method of Efficient Simulation of Physics Applications for a Quantum Computer

TL;DR

This paper introduces a hybrid simulation method combining full-state and Clifford simulators to efficiently emulate multi-qubit rotations in quantum chemistry, significantly reducing computational costs for large-scale quantum circuit simulations.

Contribution

A novel hybrid simulation approach that optimizes multi-qubit rotation emulation, enabling faster and more efficient quantum chemistry simulations on classical hardware.

Findings

Achieved a speedup of approximately 18 times for 24-qubit Hamiltonians.

Reduced computational costs by leveraging Pauli frame optimization.

Demonstrated integration into the Intel Quantum SDK for practical applications.

Abstract

Quantum chemistry and materials science are among the most promising areas for demonstrating algorithmic quantum advantage and quantum utility due to their inherent quantum mechanical nature. Still, large-scale simulations of quantum circuits are essential for determining the problem size at which quantum solutions outperform classical methods. In this work, we present a novel hybrid simulation approach, forming a hybrid of a fullstate and a Clifford simulator, specifically designed to address the computational challenges associated with the time evolution of quantum chemistry Hamiltonians. Our method focuses on the efficient emulation of multi-qubit rotations, a critical component of Trotterized Hamiltonian evolution. By optimizing the representation and execution of multi-qubit operations leveraging the Pauli frame, our approach significantly reduces the computational cost of simulating quantum circuits, enabling more efficient simulations. Beyond its impact on chemistry applications, our emulation strategy has broad implications for any computational workload that relies heavily on multi-qubit rotations. By increasing the efficiency of quantum simulations, our method facilitates more accurate and cost-effective studies of complex quantum systems. We quantify the performance improvements and computational savings for this emulation strategy, and we obtain a speedup of a factor $\approx 18$ ($\approx 22$ with MPI) for our evaluated chemistry Hamiltonians with 24 qubits. Thus, we evaluate our integration of this emulation strategy into the Intel Quantum SDK, further bridging the gap between theoretical algorithm development and practical quantum software implementations.