Treat double as float to use HW floating point

We have the hardware floating point support in our core. We are using 14.0 quartus and nios ide. We are on a cyclone iv.

We have a library that someone else has built and they use double throughout their code.

Is there a compiler setting or something that we can put in the nios ide that will treat all doubles as floats that will then take automatic advantage of our hardware acceleration for floating point?