Switching between float and double precision at compile time
Where should I look if I want to switch between float and double precision at compile time. For example, if the user wants everything in floating position instead of double precision, how can I maintain this flexibility? In other words, how do you define a variable that can be either floating or conditionally double precision?
If you make a key at compile time, a simple one typedef
will do:
#ifdef USE_DOUBLES
typedef double user_data_t;
#else
typedef float user_data_t;
#endif
Use user_data_t
in your code and install USE_DOUBLES
if you want doubles
at compile time:
g++ -DUSE_DOUBLES=1 myprogram.cpp
Without knowing exactly how you intend to use the data, it is difficult to recommend the right solution.
Look at the union date type .
http://msdn.microsoft.com/en-us/library/5dxy4b7b(v=VS.80).aspx
Templates will also be viable depending on use.
http://msdn.microsoft.com/en-us/library/y097fkab.aspx
I prefer not to have it #define
in my code.
I would have two different headers with two different typedef
s and allow build options to choose which header is included.