Towards simulating the Epoch of Reionization on exascale supercomputers — 

The Epoch of Reionization (EoR) represents a fundamental phase transition in cosmic history, demanding sophisticated numerical simulations to interpret observations of the high-z Lyman alpha forest, combined with HST/JWST, and upcoming 21cm surveys. Our latest Cosmic Dawn III simulation has redefined the state-of-the-art by simultaneously and precisely predicting multiple EoR observables: the evolution of the neutral Hydrogen fraction, Thomson scattering optical depth, ionizing photon rates, Lyman-α forest optical depth distributions, and ionizing mean free path evolution. However, critical limitations persist, including insufficient spatial resolution for internal galaxy physics and box sizes still to small for interpreting observed volumes like ASPIRE/EIGER. Breaking through these barriers requires exascale computing and new computational frameworks, which I will illustrâtes. Additionally, I will discuss the underexplored limitations of the M1 radiative transfer model, presenting new work by PhD student Mei Palanque on Pn methods coupled to photochemistry, which quantifies M1 artifacts and opens pathways to more accurate EoR modeling.