A High-Performance Parallel Computing Framework for Uncertainty Quantification Analysis of RF Devices*

We present a high-performance, integrated environment for uncertainty quantification (UQ) and reliability analysis (RA) of Vacuum Electronic RF devices based on Sandia National Lab’s DAKOTA toolkit, NRL’s GPU-accelerated, 3D Electromagnetic PIC code Neptune, and AFRL’s high-performance computing management tool, the Galaxy Simulation Builder. The environment allows for UQ studies of RF devices to be incorporated in the device design process and to be deployed, carried out, monitored and analyzed in a user-friendly and flexible fashion using a diverse selection of high-end parallel computing resources.