Floating point calculation in MWS?

Is it possible to switch between floating point and double precision in MWS? I know from my experience with other software that floating point calculation may increase the speed a lot and reduce the memory requirement. Still the accuracy of the calculations arent to bad at all.