Quantum Software for the Utility-Scale and Beyond

Abstract

With quantum computers comprised of over 100 qubits now generally accessible, a multitude of demonstrations at this utility scale have been performed. Although the focus is typically on the scale, quality, and speed of the quantum hardware, it is imperative that the classical computing infrastructure underpinning quantum computing architectures be performant and reliable, in lockstep with quantum hardware improvements. It must also be scalable to meet the eventual needs of an error-corrected quantum platform. This includes the tools and capabilities needed for the tight integration between scalable quantum and classical computing resources necessary to enable Quantum Centric Supercomputing. Moreover, as quantum platforms continue to mature, software needs to support the continued abstraction away from quantum circuits and operators, which is necessary for the widespread adoption of quantum computing as a computational resource. In this talk, Jay M. Gambetta provides an overview of the progress to date made at IBM Quantum in these areas. Through in-depth open-source benchmarking, it becomes evident that the focus on performance and reliability for Qiskit makes it the preeminent quantum software development kit for constructing, manipulating, and optimizing quantum circuits at the utility scale and beyond. This advantage is extended with the Qiskit Transpiler Service, which further improves the optimization of quantum circuits while simultaneously reducing overall runtime using reinforcement learning. Efforts to create an open and transparent quantum benchmarking platform for the community will also be discussed. For Quantum Computational Scientists working above the level of quantum assembly code, a catalog of Qiskit Functions, created by both IBM Quantum and select partners, is announced. This catalog targets several domain-specific application areas and error mitigation services, utilizing Qiskit Runtime and the managed resource capabilities of Qiskit Serverless. The advantages that these functions offer will be demonstrated through several examples. An open user execution framework allows for independent verification of these claims. Taken together, the simultaneous focus on both quantum hardware and classical software performance promises to accelerate the journey towards useful quantum computation in the coming years.