Scalable Frameworks for High-Performance Quantum Computation

Scalable Frameworks for High-Performance Quantum Computation

Our research focuses on developing scalable, scientifically validated frameworks for simulating and benchmarking quantum algorithms on modern HPC systems. Building on work conducted on NERSC’s Perlmutter system, and extending toward the next-generation Doudna system expected in 2026, we have established a benchmarking infrastructure for evaluating quantum circuits and Hamiltonian-based simulations across realistic problem domains. This enables consistent, reproducible analysis of performance and supports meaningful comparison of emerging quantum approaches.

A central objective is advancing hybrid quantum-classical workflows that integrate quantum simulation with classical computation on GPU-accelerated, multi-node systems. We are investigating locality-aware decomposition strategies that represent large quantum systems through coordinated collections of smaller computational kernels. These methods provide a pathway to improved scalability while introducing challenges in data movement, orchestration, and distributed performance.

To address these challenges, we combine detailed performance measurement with structured workflow coordination, using machine learning selectively where it improves efficiency and adaptability. This approach enables simulation workflows that respond to both problem structure and hardware characteristics, supporting more effective use of large-scale HPC resources.

In parallel, we are developing open, extensible software frameworks that emphasize reproducibility, portability, and alignment with modern HPC environments. These tools are designed to support researchers and applied teams in fields such as chemistry, materials science, and optimization, enabling evaluation of quantum methods on realistic systems without requiring specialized infrastructure development.

This work addresses key questions around the scalability and practical utility of quantum computing. By providing validated benchmarking methodologies and computational tools for hybrid quantum-classical systems, we aim to support rigorous evaluation of quantum approaches and accelerate their integration into scientific and applied research workflows.