“…As each crossing must be completed sequentially in order, the work to complete traversal of that cycle cannot be parallelized, resulting in an extremely long runtime for each power iteration. Additionally, as it is necessary to use domain decomposition when simulating large full core reactor problems on supercomputers [25,26,29,30,31] (as is discussed in subsection 2.5), traversal of full track cycles every power iteration may require that a ray is transmitted between subdomains handled by different processes many thousands of times each power iteration, which may incur impractical synchronization and network communication costs. 5.…”