FLT_EVAL_METHOD
From cppreference.com
Defined in header <float.h>


#define FLT_EVAL_METHOD /* implementation defined */ 
(since C99)  
Specifies the precision in which all floatingpoint arithmetic operations other than assignment and cast are done.
Value  Explanation 
negative values except 1

implementationdefined behavior 
1

the default precision is not known 
0

all operations and constants evaluate in the range and precision of the type used. Additionally, float_t and double_t are equivalent to float and double respectively 
1

all operations and constants evaluate in the range and precision of double. Additionally, both float_t and double_t are equivalent to double 
2

all operations and constants evaluate in the range and precision of long double. Additionally, both float_t and double_t are equivalent to long double 
[edit] Notes
Regardless of the value of FLT_EVAL_METHOD, any floatingpoint expression may be contracted, that is, calculated as if all intermediate results have infinite range and precision (unless #pragma STDC FP_CONTRACT is off)
Cast and assignment strip away any extraneous range and precision: this models the action of storing a value from an extendedprecision FPU register into a standardsized memory location.
[edit] See also
C++ documentation for FLT_EVAL_METHOD
