DescriptionSimulations in Quantum mechanics get larger and thereby more compute intensive. Hardware architectures and programming models evolve rapidly. The challenge becomes to identify the parts of the program that can benefit from parallelization on CPU or GPU offloading. Keeping in mind that development time is a limited resource in the natural sciences.
This talk will show the results of a study on the Time Dependent Schrödinger Equation. Starting with a brief explanation of the physical problem and the design of the simulation. Identifying the constraints in different parts of the simulation. The talk will also address performance gains of GPU offloading both with OPEN-MP and cuSpsarse obtained in the simulation.