Brute-forcing spin-glass problems with CUDA

Abstract We demonstrate how to compute the low energy spectrum for small ( N ? 50 ), but otherwise arbitrary, spin-glass instances using modern Graphics Processing Units or similar heterogeneous architecture. Our algorithm performs an exhaustive (i.e., brute-force) search of all possible configurations select S ? 2 lowest ones together with their corresponding energies. mainly focus on Ising model defined arbitrary graph. An open-source implementation based CUDA Fortran and a suitable Python wrapper are provided. As opposed heuristic approaches, ours is exact thus can serve as references point benchmark other algorithms hardware, including quantum digital annealers. offers unprecedented speed efficiency already visible commodity hardware. At same time, it be easily launched professional, high-end graphics cards virtually at no extra effort. practical application, we employ that recent Matrix Product State – despite its one-dimensional nature still accurately approximate fully connected graphs size approaching 50.