Quantum Sampling Reaches a Nucleus Beyond a 768-GB Node's Memory
An IIT Roorkee team produced an unconverged approximate ground-state energy for the exotic nucleus magnesium-32 by sampling configurations from IBM quantum processors after their 768-gigabyte supercomputer node ran out of memory, the first use of sample-based quantum diagonalization on the nuclear shell model. A close look at why nuclear structure is a defense workload, how it feeds stockpile stewardship simulation and evaluated nuclear data, and the accuracy gap still separating a rented quantum sampler from a production weapons code.
A Quantum Fluid Solver Gets an Exascale Reality Check
An Oak Ridge team ran a variational quantum linear solver for fluid dynamics end to end on the Frontier exascale machine and published the honest bill: memory costs cut over a thousandfold, circuit design rules that fail to predict convergence, and every qubit still emulated. For the discipline behind hypersonic flow simulation, it is the first production-scale accounting of what a quantum CFD pipeline actually costs.