HIGH RMSE value

do i need to use hyperparameter tuning in the decision tree regression at the starting or I should use it where it is asked?
I am getting a high rmse value in the validation by not using the tuning.