Sometimes, the optimization procedure takes 5 or 6 days time to find its minimum optimized structure in Gaussian 03W. Is there any procedure to speed up the optimization process for large sized molecules?
For large sized molecules you could preoptimize the structure in a low-level calculation, e.g. by using smaller basis sets. Moreover, you could try the 'convergence related options' offered by the code.
First optimize your structure at semi-classical level (it takes some minutes). Then use the optimized structure as input for DFT (or higher level ) optimization. This procedure reduces the optimization time considerably.