This study presents a high-performance parallel CFD platform designed for two-way coupled six-degree-of-freedom (6-DOF) fluid-structure interaction (FSI) involving large rigid-body motions. To overcome the limitations of grid distortion and remeshing costs inherent in body-fitted approaches, the platform integrates a Building-Cube Method (BCM)[1] hierarchical Cartesian mesh with a sharp-interface immersed boundary method (IBM)[2-3]. This combination allows for the simulation of complex moving geometries on a fixed background grid while ensuring rigorous boundary treatment. Geometry handling relies on the automated processing of closed STL surfaces. In cut cells, interface reconstruction facilitates precise wall-boundary enforcement and load evaluation. Hydrodynamic forces and moments, integrated from pressure and viscous tractions, directly drive a strongly coupled 6-DOF solver formulated in a single global inertial frame. This approach eliminates the cumulative errors and complexity associated with frequent local-frame transformations. To streamline pre-processing for multi-body configurations, mass properties and the initial inertia tensor are computed automatically from STL polyhedra via divergence-theorem-based surface integrals. For massively parallel performance, the solver employs cube-based domain decomposition ordered by Morton Z-order space-filling curves to optimize data locality and minimize MPI communication. The platform is validated against canonical 6-DOF cases, including flow-induced rotation and self-propelled motion. Results demonstrate the accurate reproduction of time-dependent loads, attitude evolution, and wake organization. In conclusion, the proposed BCM-IBM-6DOF framework offers a robust and scalable solution for large-motion FSI applications in free-flight dynamics and bio-inspired propulsion.