Skip to Main Content
The solution of large and complex coupled electromechanical problems requires high-performance computing resources. Over the past years, the use of graphic processing units (GPUs) in scientific computing has gained increasing popularity because of their low cost and parallel architecture. In this paper, the authors report the main results of a GPU approach for the parallelization of a research code for electromagnetic launcher analysis. Programming a GPU-based environment poses a number of critical issues that have to be carefully addressed in order to fully exploit system potential. Data have to be properly organized in order to fit the single-instruction multiple-data scheme; data transfer between the host and the device, as well as memory management of the GPU, deserves accurate programming. Two examples of application of the parallelized code have been reported to show the performance improvements that can be obtained in the numerical analysis of both rail and induction launchers.