An efficient, high-fidelity numerical aerodynamic shape optimization tool is presented. The algorithm includes an integrated geometry parameterization and mesh movement scheme based on B-spline volumes, an efficient parallel Newton-Krylov-Schur algorithm for solving the three-dimensional Reynolds-Averaged Navier-Stokes (RANS) equations, a discrete-adjoint gradient evaluation, and a gradient-based optimizer which is capable of performing large-scale optimizations subject to linear and nonlinear constraints. Several cases are presented to demonstrate the performance of the algorithm. First, an optimization is performed for a rectangular wing that is initially fit with NACA0012 sections in order to demonstrate the robustness of the mesh movement and flow analysis given substantial changes in the geometry. The optimizer is able to achieve substantial drag reduction at the target lift by altering the camber and by increasing the sweep angle. We then present a study of the wing geometry extracted from the Common Research Model (CRM) wing-body geometry; we consider the CRM wing with a sharp trailing edge, as well as a wing with the same planform, but given NACA0012 sections. Given section and twist design variables, each initial geometry yields an optimized design that demonstrates improved drag compared to the initial shape. The optimizations of the planar wing with NACA0012 sections and the CRM wings were additionally run with an Euler-based algorithm; RANS analyses were performed on the Euler-optimized geometries such that they could be compared directly with the results of the RANS-based optimizations. In the case of the planar wing with NACA0012 sections, which specified a low target lift coefficient, the Euler-based optimizer produced a very similar design which yielded the same drag coefficient as the RANS-based optimization. However, the CRM study shows that the RANS-based optimizations result in designs with much lower drag compared to the Eulerbased optimizations. We conclude that, in general, viscous and turbulent effects should be taken into account when performing aerodynamic shape optimization.