If the destination database is newly created (Run 1) performance is slower as the database has to be extended for the new data. Subsequent runs (Run 2) into the same database are generally faster.