Formal verification of floating-point arithmetic remains challenging due to non-linear arithmetic behavior and the tight coupling between control and datapath logic. Existing approaches often rely on high-level C models for equivalence checking against Register Transfer Level (RTL) designs, but this introduces abstraction gaps, translation overhead, and limits scalability at the RTL level. To address these challenges, this paper presents a scalable methodology for verifying floating-point arithm...