Which one is not an advantage of stochastic gradient descent algorithm over gradient descent algorithm?

a$.$It is applicable even if not all data fits into main memory or is available at the same time

b$.$It is more computationally efficient for large data sets

c$.$It is insensitive to the initial values