I wanted to check how much time it might take to inverse a 205 by 205 matrix in GAUSS. I am estimating a model using maximum likelihood technique with about 205 parameters and trying to inverse the 2nd derivative of the loglikelhood function to get the standard errors. It seems that this step is taking more time compared to the model convergence. I just wanted to check whether it is reasonable and also if there is any way to reduce the amount of time while calculating the standard error using inverse of the Hessian.
Computing a matrix inverse is a time consuming step, so even with an optimized version of the algorithm it is likely to take up quite a bit of time relative to other operations in a program. Someone else will comment on the second part of your question.
A bootstrap might be faster if convergence is really very fast relative to inversion. With 100 estimations, standard deviations of each parameter could be generated from the re-samples results.