Comparisons of efficiency and power of the early steam engines is difficult for several reasons: 1) there was no standard weight for a bushel of coal, which could be anywhere from 82 to 96 pounds. 2) There was no standard heating value for coal, and probably no way to measure heating value. The coals had much higher heating value than today's steam coals, with 13,500 BTU/pound sometimes mentioned. 3) Efficiency was reported as "duty", meaning how many foot pounds of work lifting water were produced, but the mechanical pumping efficiency is not known. 
Computational models for simulating combustion and heat release rates of HCCI engines require detailed chemistry models.   This is largely because ignition is more sensitive to chemical kinetics than to turbulence/spray or spark processes as are typical in SI and diesel engines. Computational models have demonstrated the importance of accounting for the fact that the in-cylinder mixture is actually in-homogeneous, particularly in terms of temperature. This in-homogeneity is driven by turbulent mixing and heat transfer from the combustion chamber walls. The amount of temperature stratification dictates the rate of heat release and thus tendency to knock.  This limits the usefulness of considering the in-cylinder mixture as a single zone, resulting in the integration of 3D computational fluid dynamics codes such as Los Alamos National Laboratory's KIVA CFD code and faster solving probability density function modelling codes.