I am building an application where two instances, running on two seperate machines, must make the same floating point calculations given the same inputs.
I know that there can be problems with double values. The two computers will not always make the same calculation due to differrences in precision settings in the FPU. FPU settings can be changed at any time by other programs that are running concurrently.
Does the same apply to single values? or, are single values at a low enough precision, that even the lowest precision FPU settings will not affect the consistency of calculations on seperate computers?
I am looking for a simple Yes or NO answer with a good explanation of why.