The Flamelet Generated Manifold (FGM) method is a promising technique in engine combustion modeling to include tabulated chemistry. Different methodologies can be used for the generation of the manifold. Two approaches, based on igniting counterflow diffusion flamelets (ICDF) and homogeneous reactors (HR) are implemented and compared with Engine Combustion Network (ECN) experimental database for the baseline n-heptane case. Before analyzing the combustion results, the spray model is optimized after performing a sensitivity study with respect to turbulence models, cell sizes and time steps. The standard High Reynolds (Re) k-ε model leads to the best match of all turbulence models with the experimental data. For the convergence of the mixture fraction field an appropriate cell size is found to be smaller than that for an adequate spray penetration length which appears to be less influenced by the cell size. With the optimized settings, auto-ignition and flame lift-off length are analyzed. In general, both techniques capture the qualitative trend of experimental results. However, typically, the HR tabulation method predicts shorter ignition delay and LOL results than the ICDF method. In a quantitative sense, the ICDF and HR methods give better results in LOL and auto-ignition predictions, respectively.