Why Quantum Needs HPC
Key insight: quantum computing complements HPC, it does not replace it. Even a fully quantum-native application still needs a classical supercomputer around it, for reasons that don’t go away as hardware improves:
- Classical control electronics — every quantum computer is driven by extensive classical infrastructure (see What is a Quantum Computer); the QPU is never standalone.
- Classical optimization loops — algorithms like VQE are fundamentally iterative classical-quantum loops, not one-shot quantum computations.
- Pre- and post-processing — even the most quantum-native applications need classical work before (problem encoding, basis sets, qubit mapping) and after (result interpretation, error mitigation post-processing) the quantum part runs.
- Most of the workload stays classical — QPUs target narrow, specialized kernels within a much larger computation; the surrounding orchestration, data movement, and analysis is HPC’s job.
Reframing the workflow
This is the same Map → Optimize → Execute → Post-process pattern from earlier in this vault, just viewed at HPC scale: classical preprocessing on HPC → problem decomposition/encoding → QPU execution of a specialized kernel → HPC postprocessing. The steps don’t change; what changes is that “classical” now means a full supercomputer, not a laptop.
Related
- What is HPC
- Qiskit Patterns — the same four-step pattern, at circuit scale
- Variational Quantum Eigensolver (VQE)
- Quantum-Centric Supercomputing (QCSC)
Self-Check
- Could you explain why “quantum computing complements HPC” rather than replaces it, using VQE as an example?
- Name the four reasons a quantum computer still needs HPC around it, even hypothetically at its best.
- How does this note’s four-step workflow map onto Qiskit Patterns’s Map → Optimize → Execute → Post-process?