Atomistic simulations have become a valuable tool for studying crack propagation, revealing mechanisms of crack growth with nanometer-scale precision and picosecond-scale resolution. While these simulations advance our understanding of material failure, a challenge persists in bridging atomistic models with experimental observations, even in brittle materials where the material separation process is localized to the crack tip. This dissertation is composed of two papers and related unpublished work addressing this challenge through a systematic investigation of subcritical crack growth and arrest in silica glass.First, a novel approach was developed that is both computationally efficient and robust to noise, enabling accurate inference of crack-tip location and associated stress intensity factors (SIFs) from large, noisy molecular dynamics datasets. Building on this capability and leveraging advanced parallel computing, long-timescale molecular dynamics simulations were performed to probe crack growth rates well below the dynamic limit. By employing carefully selected geometries and boundary conditions, growth rates were quantified as a function of the crack driving force, enabling direct comparison with experimental measurements. The correspondence between simulations and experiments suggests that atomistic modeling is capturing the underlying mechanism of subcritical crack growth, specifically a thermally activated stick–slip process that persists up to crack velocities approaching 1 km/s. In addition to crack growth dynamics, the structural and mechanical environments around arrested cracks were examined. Microstructural variations in density, bond counts, coordination defects, and stress fields were evaluated to assess their potential role in resisting crack advance. While certain features recur near arrest sites, no single structural or stress based indicator was found to be sufficient to predict arrest deterministically, highlighting the possible combined influence of microstructural heterogeneity, evolving stress state, and stochastic fluctuations. Together, these papers and further studies establish a foundation for understanding subcritical crack growth in silica glass through atomistic simulations. They demonstrate how atomistic modeling can bridge nanoscale fracture mechanisms with experimental observations, offering insight into the fundamental processes that control crack advance and arrest in brittle materials.