Quantum-centric supercomputing: a new computational framework for chemistry

Over the past decades, algorithms for classical computers have revolutionized the simulation of chemistry, bringing new insights into the properties and behavior of molecules. Today, we are at an inflection point in the evolution of quantum computing, where quantum systems have reached the scale and quality needed to tackle challenging scientific problems beyond the reach of exact brute-force classical methods. To sustain and expand this research, we believe that existing classical workf lows should be integrated with quantum components and executed on heterogeneous quantum-classical platforms, in a computational framework that we call quantum-centric supercomputing (QCSC). Within QCSC algorithms and architecture, quantum-computing components will deliver a value that cannot be matched by classical or quantum computations alone. With QCSC,researchers will be supported in making new algorithmic discoveries. One such advance is the sample-based quantum diagonalization method, which has made electronic structure calculations previously considered years away feasible on present-day quantum devices. This method was used to simulate nitrogen triple-bond dissociation (N2) in a correlation-consistent cc-pVDZ basis and the active-space electronic structure of methyl-capped [2Fe–2S] and [4Fe–4S] clusters. The latter system is beyond the scale (i.e. number of electrons and orbitals) of current exact diagonalization calculations. We are committed to facilitate and accelerate algorithmic discovery by the general availability of quantum devices and software tools. We believe that QCSC is the most natural path to deliver quantum advantage– and the foundation of the quantum era to come.