Although recent experimental progress in near-term quantum computers has generated significant interest and optimism, a large gap remains from these quantum computers to practical applications with real-world value. Much of this gap can be attributed to the current design of quantum system stacks, which impede optimizations bridging from the quantum computer science to the quantum device physics. In particular, current system architectures prevent software from optimally compiling quantum programs or performing low- level error mitigation. This proposal outlines software development that will enable optimizations that cross between layers of the stack, unlocking ~20x efficiency gains that were previously invisible. The proposed software will have two objectives. Objective A, decomposition to pulse-level controls, will enable quantum programs to be optimally expressed in terms of the native capabilities of the underlying quantum hardware. Objective B, low-level error mitigation, will enable the compiler to integrate noise suppression techniques with traditional compilation objectives such as gate cancellation and crosstalk avoidance. The work plan is divided into 7 Tasks, beginning with the development of software tools for interfacing with a wide range of qubit (quantum bit) technologies. Collaboration is anticipated with national research testbeds, as well as industry leaders in quantum hardware. Tasks 2-4 and 5-7 will fulfill Objectives A and B respectively; they are scoped for completion by November 2021, which will enable successfully transition to Phase II. The open source software developed under this proposal will demonstrate the feasibility of a quantum software stack that enables cross-layer optimizations for significant efficiency gains. The market for quantum computing is growing rapidly, and the software segment exhibits particularly fast growth with an estimated 64% year-over-year growth forecasted. Successful execution of Phase I will enable capture of a large fraction of that market, by enabling software optimizations that make practical applications achievable 5 years sooner than otherwise possible. These practical applications will include tasks that are intractable for classical computers. For example, ongoing research with utility companies has suggested quantum speedups for solving power grid optimization tasks that are otherwise intractable. In the broader Phase II and Phase III horizon, quantum applications will manifest in sectors important to national security, such as improving aircraft design. There are also anticipated public benefits through improved molecular simulation for tasks like higher- efficiency fertilizer production, which currently consumes 3% of the worlds total energy output. The proposed technology will accelerate these commercial applications and benefits.
