Intrinsically disordered proteins are characterized by their large manifold of thermally accessible conformations and their related statistical weights making them an interesting target of simulation studies. To assess the development of a computational framework for modeling this distinct class of proteins, this work examines temperature-based replica exchange simulations to generate a conformational ensemble of a 28-residue peptide from the Ebola virus protein VP35. Starting from a prefolded helix-β-turn-helix topology observed in a crystallographic assembly, the simulation strategy tested is the recently refined CHARMM36m force field combined with a generalized Born solvent model. A comparison is provided of two replica exchange methods where one is a traditional approach of a fixed set of temperatures and the other is an adaptive scheme of allowing the thermal windows to move in temperature space. The assessment is further extended to include a comparison to equivalent CHARMM22 simulation datasets. The analysis finds CHARMM36m to shift the minimum in the potential of mean force (PMF) to a lower fractional helicity as compared to CHARMM22, while the later showed greater conformational plasticity along the helix-forming reaction coordinate. Among the simulation models, only the adaptive tempering method with CHARMM36m found an ensemble of conformational heterogeneity consisting of transitions between -helix-β-hairpin folds and unstructured states that produced a PMF of fractional fold propensity in qualitative agreement with circular dichroism experiments of reporting a disordered peptide.