hi is there an option to treat all float variables (32bit) to double (64bit) on a 64bit machine? I found -mdouble-float, but I got error message cc1: error: unrecognized command line option "-mdouble-float" not sure why. appreciated if anyone can commend. thanks Qianqian