C++ Library - <cfenv>

Previous
Quiz
Next

The <cfenv> header in C++, provides various functions to deal the floating-point environment, along with floating-point rounding modes and handling exceptions.

The floating-point environment maintains a series of status flags and control modes that indicate if any specific floating-point errors have occurred. The Rounding mode refers to how floating-point numbers are rounded during arithmetic operations.

Including <cfenv> Header

To use the cfenv in C++, we need to include the cfenv header file as shown below.

#include <cfenv>

Functions of <cfenv> Header

Below is list of all functions from <cfenv> header.

Rounding Direction Functions

The <cfenv> header provides some rounding modes by using those we can control the rounding direction for floating-point operations.

S.NO Functions & Description
1

fegetround(void)

This function returns a value that indicates the rounding direction mode in the current floating point environment.
2 fesetround(int rdir)

Sets rdir as the current rounding direction mode for the floating point environment.

Getting the Current Rounding Mode

In the following example we are going to use fegetround() to get the current rounding direction mode of the floating-point environment.

#include <iostream>
#include <cfenv>
int main() {
   std::fesetround(FE_DOWNWARD);  
   int round_mode = std::fegetround();  
   if (round_mode == FE_DOWNWARD)
      std::cout << "Rounding mode: Downward" << std::endl;
   else
      std::cout << "Other rounding mode" << std::endl;
}

Output

If we run the above code it will generate the following output 

Rounding mode: Downward

Floating-point exceptions

In C++ Floating-point exceptions when a floating-point operation leads to an error condition. such as division by zero or overflow. some common floating-point exceptions are as follows.

Floating-point exceptions

In C++, floating-point exceptions occur when an arithmetic operation results in an error condition, such as division by zero or overflow. Some common floating-point exception functions are as follows:

S.NO Functions & Description
1 feclearexcept()

Clears the specified floating-point exceptions.
2 feraiseexcept()

Raises the specified floating-point exceptions.
3 fegetexceptflag()

Stores the current state of specified floating-point exceptions in the given object.
4 fesetexceptflag()

Sets the floating-point exceptions as indicated by the state stored in the object.

Clearing a Floating-point Exception

In the following example we are going to use feclearexcept() to clear specific floating-point exception.

#include <iostream>
#include <cfenv>

#pragma STDC FENV_ACCESS ON
int main() {
   std::feclearexcept(FE_DIVBYZERO);  
   volatile double result = 1.0 / 0.0;
   if (std::fetestexcept(FE_DIVBYZERO))
       std::cout << "Division by zero exception detected"  << std::endl;
   else
      std::cout  << "No exception"  << std::endl;
}

Output

If we run the above code it will generate the following output 

Division by zero exception detected

Manipulating Entire Floating Point Environment

C++ provides functions to save, restore, or modify the entire environment at once. The floating-point environment includes both the current rounding mode and exception flags.

S.NO Functions & Description
1 fegetenv()

This function attempts to store the current state of the floating-point environment in the object.
2 fesetenv()

This function attempts to establish the state of the floating-point environment as represented by the object.
3 feholdexcept()

This function Saves the current state of the floating-point environment in the object.
4 feholdexcept()

This function attempts to establish the state of the floating-point environment as represented by the object.
5 fetestexcept()

This function returns the exceptions currently set, among those specified by excepts.

Saving and Restoring the Environment

In the following example we are going to use feholdexcept() to save the current floating-point environment both rounding modes and exception flags and clears the exception flags.

#include <iostream>
#include <cfenv>
#pragma STDC FENV_ACCESS ON
int main() {
   fenv_t env;
   std::feholdexcept(&env);  
   volatile double result = 1.0 / 0.0;
   if (std::fetestexcept(FE_DIVBYZERO))
      std::cout << "Division by zero detected!" << std::endl;
   std::fesetenv(&env);
}

Output

If we run the above code it will generate the following output 

Division by zero detected!
Print Page
Previous
Next
Advertisements