Based on the various texts I've browsed, the term "advanced calculus" can refer to anything beyond a one-year "mechanical" (i.e., the focus is computational facility rather than theoretical understanding) course in one-variable calculus. Thus "advanced calculus" could refer to (A) one-variable real analysis, (B) a computational course dealing with partial derivative, multiple integrals, and vector calculus, (C) a mathematically rigorous course on multivariable calculus (generally using differential forms), or (D) a first course in complex variables or Lebesgue theory. Thus the term is incredibly vague. On one extreme, I've seen an advanced calculus book with nary a proof. (For example, see the Schaum's study guide on advanced calculus.) On the other extreme, Loomis and Sternberg covers calculus in the setting of Banach spaces, and introduces a large amount of abstract, linear, and multilinear algebra, in order to rigorously describe differentiable manifolds embedded in Banach spaces.
As far as I'm aware, there are really only three undergrad books involving "calculus on manifolds": Spivak's Calculus on Manifolds, Munkres's Analysis on Manifolds, and Hubbard and Hubbard's Vector Calculus, Linear Algebra, and Differential Forms. In order of increasing difficulty, they are Hubbard and Hubbard, Munkres, and Spivak. Each of them really only talk about manifolds in R^n, although Munkres, and especially Spivak give clues as to how to generalize them. Hubbard and Hubbard hide some of the more difficult proofs out of gentleness, while Spivak hides some proofs and writes some other very unclear, opaque, or just plain wrong ones out of lack of pedagogical experience (Calculus on Manifolds is his first book, written when he was 25). Thus, the best one out of the three is probably Munkres, who goes out of his way to help students develop intuition. Hilariously, Loomis and Sternberg was written for freshmen (at Harvard...), but its difficulty makes it suitable for few undergrads. It is a beautiful book, but really more appropriate for grad students (but then again, they insist on using Jordan measure instead of Lebesgue in the name of keeping things elementary...)