The HLRB cluster as quantum CISC compiler

The project encompasses matrix method developments, tailored parallelization as well as cutting-edge applications exploiting the power of the HLRB-II cluster: fast matrix exponential algorithms using Chebyshev series are devised in view of calculating quantum dynamics of large systems. They outperform the standard Padé-approximation by a speed-up of approximately 30% in CPU time while obtaining even better accuracy. The routines are incorporated into a fully parallelized package of gradient-flow algorithms for optimal quantum control. As an application, here we present a quantum CISC compiler: it breaks large target unitary gates into modules of effective m-qubit (i.e. two-level system) interactions. We extend the standard restricted set of modules with m = 1,2 (RISC) to a scalable toolbox of multi-qubit optimal controls with m ≤ 10 forming modules of complex instruction sets (CISC). Typically, the instruction code ('experimental controls') by our quantum CISC compiler is some three to ten times faster than by RISC compilation thus dramatically saving the essential quantum coherences from unnecessary relaxative decay with time. This advantage of our method over standard universal gates is demonstrated for the indirect SWAP gate, the quantum Fourier transform as well as for multiply-controlled NOT gates.