In tensorflow what is the difference between trainable and stop gradient

I would like to know the difference between the option trainable=False and the tf.stop_gradient(). If I make the trainable option False will my optimizer not consider the variable for training? Does this option make the it a constant value throughout the training?