Architecture and Design of a Small Parallel Supercomputer for Modelling Bolometric Lightcurves of Type II-P Supernovae

Processing raw luminosity of supernovae into lightcurves across a range of observed and unobserved wavelengths requires integration across individual wavelength windows to establish luminosity over a broad range. Martinez et al. (in review, 2021) detail a variety of methods for bolometric lightcurve computation. Different portions of the curve may be modelled simultaneously, making parallel computing potentially useful for lightcurve computation. We have built a six-node small parallel supercomputer (SPS@UCA) to accommodate the bolometric lightcurve computation program SuperBoL (Lusk and Baron, 2017) and future computationally-intensive projects. We present its architecture, design, and specifications, its uses for both computational astronomy and outreach, and a discussion of ongoing modelling.