As of now, it looks like they're ignoring it: https://github.com/ran-j/PS2Recomp/blob/91678d19778891b4df854af0a2e9d5c162fb42c1/ps2xRuntime/output/ps2_runtime_m

As of now, it looks like they're ignoring it: https://github.com/ran-j/PS2Recomp/blob/91678d19778891b4df854af0a2e9d5c162fb42c1/ps2xRuntime/output/ps2_runtime_macros.h#L133 https://github.com/ran-j/PS2Recomp/blob/91678d19778891b4df85... #define FPU_ADD_S(a, b) ((float)(a) + (float)(b)) (etc) But if you wanted to handle it, you'd presumably macro expand the floating point operations to something that matches the PS2 fpu (or comes closer).