Quantum Computer Programming, Compilation, and Execution with XACC

We demonstrate how near-term quantum processing units (QPUs) can be integrated into high-performance computing using applications for quantum chemistry, machine learning, and combinatorial optimization. The eXtreme-scale ACCelerator programming model (XACC) is an open-source framework that supports quantum acceleration of scientific workflows across many different vendor QPUs. We develop the XACC programming model as a coprocessor model akin to the design of OpenCL or CUDA for GPUs, in which the framework offloads computational work by defining quantum kernels for execution on an attached QPU accelerator. We demonstrate an extensible quantum compilation mechanism with general quantum circuit optimization and transformation capabilities. We show how this approach is agnostic to quantum programming language and QPU hardware, and we demonstrate how XACC enables hybrid computing programs to be ported to multiple processors for benchmarking, verification and validation. Finally, we measure the utility of this programming model by demonstrating a distributed-memory implementation of the variational quantum eigensolver.