This work will focus on assessing the performance of the basic compute kernels of deterministic particle transport codes on heterogeneous computing resources. The goal is to understand what penalties and benefits to expect and understand if algorithmic changes to the base solver algorithm can be made to eliminate poor performance.