A Full Stack Framework for High Performance Quantum-Classical Computing

TL;DR

This work tackles the challenge of integrating quantum processing units (QPUs) into scalable HPC environments to enable large-scale hybrid quantum-classical workloads. It presents a full-stack, device-agnostic architecture comprising a Quantum Interface Library, Adaptive Circuit Knitting, and Quantum Compiler Extension, all supported by Cray LLVM and QIR to target diverse backends. The authors demonstrate three workloads—a HHL-based quantum linear solver, a QAOA2-based MaxCut workflow, and adaptive circuit knitting on multi-GPU/Multi-QPU scales—validating functionality and performance on HPE Cray EX systems and at ISC24. By leveraging existing HPC infrastructure (MPI, Slurm, Cray libraries) for distributed orchestration and retargetable compilation, this framework offers a practical path to unified quantum-classical programming within classical HPC stacks and paves the way for heterogeneous quantum accelerators at exascale.

Abstract

To address the growing needs for scalable High Performance Computing (HPC) and Quantum Computing (QC) integration, we present our HPC-QC full stack framework and its hybrid workload development capability with modular hardware/device-agnostic software integration approach. The latest development in extensible interfaces for quantum programming, dispatching, and compilation within existing mature HPC programming environment are demonstrated. Our HPC-QC full stack enables high-level, portable invocation of quantum kernels from commercial quantum SDKs within HPC meta-program in compiled languages (C/C++ and Fortran) as well as Python through a quantum programming interface library extension. An adaptive circuit knitting hypervisor is being developed to partition large quantum circuits into sub-circuits that fit on smaller noisy quantum devices and classical simulators. At the lower-level, we leverage Cray LLVM-based compilation framework to transform and consume LLVM IR and Quantum IR (QIR) from commercial quantum software frontends in a retargetable fashion to different hardware architectures. Several hybrid HPC-QC multi-node multi-CPU and GPU workloads (including solving linear system of equations, quantum optimization, and simulating quantum phase transitions) have been demonstrated on HPE EX supercomputers to illustrate functionality and execution viability for all three components developed so far. This work provides the framework for a unified quantum-classical programming environment built upon classical HPC software stack (compilers, libraries, parallel runtime and process scheduling).

Figures (8)

• Figure 1: A full-stack architecture highlighting extensions to the HPC programming environment for a quantum API, workload distribution hypervisor (ACK), and quantum runtime compilation.
• Figure 2: Quantum interface library replaces circuit simulation with API calls to enable invocation of quantum kernel from classical HPC program.
• Figure 3: Hybrid MPI enabled Quantum-HPC application scheme.
• Figure 4: Monolithic hybrid jobs are split into basic blocks for optimized SLURM scheduling. The indices in J_i_j indicate the affiliation to the original job i and time ordering j. The classical resources are typically larger than the quantum resources.
• Figure 5: Hybrid quantum linear solver run log and results comparison with classical solution from BLAS (top). Schematic SLURM job status query (bottom).