FLT_ROUNDS
From cppreference.com
                    
                                        
                    
                    
                                                            
                    | Defined in header  <cfloat> | ||
| #define FLT_ROUNDS /* implementation defined */ | ||
Specifies the current rounding direction of floating-point arithmetic operations.
| Value | Explanation | 
| -1 | the default rounding direction is not known | 
| 0 | toward zero; same meaning as FE_TOWARDZERO | 
| 1 | to nearest; same meaning as FE_TONEAREST | 
| 2 | towards positive infinity; same meaning as FE_UPWARD | 
| 3 | towards negative infinity; same meaning as FE_DOWNWARD | 
| other values | implementation-defined behavior | 
Notes
The rounding mode can be changed with std::fesetround and FLT_ROUNDS reflects that change.
The possible values of FLT_ROUNDS match the possible values of std::float_round_style, returned by std::numeric_limits::round_style.
See also
| indicates floating-point rounding modes (enum) | |
| (C++11)(C++11) | gets or sets rounding  direction (function) | 
| floating-point rounding direction (macro constant) |