ieee_get_rounding_mode

Date:: 10-12-2011

NAME

IEEE_GET_ROUNDING_MODE - Returns current floating-point rounding mode

SYNOPSIS

CALL IEEE_GET_ROUNDING_MODE ([ROUND_VALUE=]round_value)
   TYPE(IEEE_ROUND_TYPE) :: round_value

IMPLEMENTATION

Cray Linux Environment (CLE)

STANDARDS

Fortran

IEEE Standard for Binary Floating-point Arithmetic

DESCRIPTION

The IEEE_GET_ROUNDING_MODE intrinsic subroutine is typically called upon entry to a procedure to obtain the current floating-point rounding mode. It accepts the following argument:

round_value: Must be scalar and of type TYPE(IEEE_FLAG_TYPE). It is an INTENT(OUT) argument.

The name of this intrinsic cannot be passed as an argument.

NOTES

The IEEE intrinsic procedures use the named constants contained in a system module, so you must include the following statement in your program:

USE, INTRINSIC :: IEEE_ARITHMETIC

RETURN VALUES

round_value is set to one of the following values. These named constants indicate the possible floating-point rounding modes.

IEEE_NEAREST
IEEE_UP
IEEE_TO_ZERO
IEEE_DOWN

EXAMPLES

Example 1:

USE, INTRINSIC :: IEEE_ARITHMETIC
REAL  xdef
REAL(8)  x8
LOGICAL x_round_near_supported
TYPE(IEEE_ROUND_TYPE) :: round_nearest_flag = IEEE_NEAREST
TYPE(IEEE_ROUND_TYPE) :: current_round_status
 x_round_near_supported = IEEE_SUPPORT_ROUNDING(round_nearest_flag,xdef)

IF (.NOT. x_round_near_supported) STOP "RND_NEAR_UNS"

CALL IEEE_GET_ROUNDING_MODE(round_status)

IF ( round_status == IEEE_NEAREST) then
      do calculations expecting round_to_nearest
ENDIF