IEEE_EXCEPTIONS Intrinsic Module

The IEEE_EXCEPTIONS module contains derived data types that include named constants for controlling the level of support, and intrinsic module procedures.

The derived types in the intrinsic modules have components that are private. The IEEE_EXCEPTIONS intrinsic module supports the setting, clearing, saving, restoring, or testing of exception flags. It defines the following derived types:

• IEEE_FLAG_TYPE: Identifies an exception flag for errors that occur during an IEEE arithmetic operation or assignment. Its values are the following named constants:

  IEEE_INVALID	IEEE_DIVIDE_BY_ZERO
  IEEE_OVERFLOW	IEEE_INEXACT
  IEEE_UNDERFLOW

  Each of the above exceptions has a flag whose value is either quiet or signaling. The initial value is quiet and it signals when the associated exception occurs. To determine the value of a flag, use intrinsic module subroutine IEEE_GET_FLAG. To change the status for a flag, use intrinsic module subroutine IEEE_SET_FLAG or IEEE_SET_STATUS.

  If a flag is signaling on entry to a procedure, the processor sets it to quiet on entry and restores it to signaling on return.

  If a flag is quiet on entry to a procedure with access to modules IEEE_ARITHMETIC or IEEE_EXCEPTIONS, and is signaling on return, the processor will not restore it to quiet.

  The IEEE_FLAG_TYPE module also defines the following named array constants:

  • IEEE_USUAL=(/IEEE_OVERFLOW,IEEE_DIVIDE_BY_ZERO, IEEE_INVALID/)

  • IEEE_ALL=(/IEEE_USUAL,IEEE_UNDERFLOW,IEEE_INEXACT/)

• IEEE_STATUS_TYPE: Saves the current floating-point status.

The IEEE_ARITHMETIC module includes support for IEEE_EXCEPTIONS module, and public entities in IEEE_EXCEPTIONS module are also public in the IEEE_ARITHMETIC module.

To see a summary of all the IEEE_EXCEPTIONS intrinsic procedures, see IEEE Intrinsic Modules Quick Reference Tables