The genome is shaped by evolution through complex, interacting forces, and so understanding the interplay of these forces often falls to fine scale -- population-level -- simulation. As the scope of genomic data increases with technological advances, so too must the scope of the simulations. However, these simulations do not scale well to parallel architectures due to inherent data dependency issues. In this talk, I will explore methods and their tradeoffs for parallelizing population-level simulations, with specific interest in leveraging heterogeneous HPC architectures.