HPC-QC integration for neutral atoms based QPUs


Giving an introduction to neutral atoms based, digital-analog, quantum computing and the challenges that arise trying to integrate such non-solid state devices into current and future HPC centres.

In this talk my main message was the importance of realising even though we might be exiting the NISQ area strictly speaking we don’t know how long until the next area might have usable applications. That is, given 10 000 qubits, we don’t know if it will be more efficient running NISQ like algorithms at a scale or fault tolerant ones. From Pasqal we have an extra stake here of course as we’re currently pushing this “digital-analog” or “analog-digital” capability (using local addressing in the Rydberg basis) where we certainly won’t be doing fault tolerant things just yet (although the Neutral atoms / Rydberg platform is well suited for this too).

Such vendor specific capabilities, and hardware-near programming, should be taken more into account when designing HPC-QC integration software stacks, in my personal opinion. Moreover, much early effort on HPC-QC integration appears to be centred around superconducting qubit hardware. When dealing with other hardware plans for what to do should take into account the specificities for that hardware rather than blindly follow what’s already been done. Shot rate, performance, usage but also maintenance and calibration can differ greatly especially between the solid-state devices and those built around e.g. atoms.