I have to design a FFT Core, which has to support floating point operations. I already tried the Xilinx CORE Generator. But it seems that The FFT Core only support 2's complement numbers. It does include a Block floating point for scaling, but is not good enough in my case ..... :?:
I'm not aware of any commercial float FFT core.(It may exist of course). Using individual scaled numbers rather than usual block scaling would reduce FFT performance considerably, that may be a reason why it isn't generally used.