ld Manual

Introduction to ld
ld Manual—529650.001
1-15
Linking Floating-Point Programs
Linking Floating-Point Programs
When linking object files using ld, you can specify the floating-point type of the output
object file using the option -set with the attribute-name FLOATTYPE.
The floating-point type can be any of these:
•
TANDEM_FLOAT
•
IEEE_FLOAT
•
NEUTRAL_FLOAT
If you do not specify the floating-point type of the output object file, ld derives it from
the types of the input files. See Link-Time Consistency Checking for more information.
When modifying an existing object file, ld sets the type as specified by the option
-change.
For more information about using IEEE floating-point format:
Link-Time Consistency Checking
ld checks the consistency of floating-point type combinations when linking object files.
The type of checking depends on whether the floattype attribute is specified:
•
floattype Not Specified
•
floattype Specified
floattype Not Specified
As summarized in Table 1-1 on page 1-16, when the floattype attribute is not
explicitly set with the -set option, ld uses the floattype attribute values of all the
input object files to determine the floattype attribute value of the output object file. If
the consistency checks of the input object files result in an invalid floating-point type or
an inconsistent value, ld issues an error message and does not create an output
object file.
Topics Sources
Building programs using IEEE floating-point
format
Guardian Programmer’s Guide
Operating mode routines
Routines for converting between Tandem
and IEEE floating-point formats. Note that
‘Tandem’ float and ‘TNS’ float are used
interchangeably.
Guardian Procedure Calls Reference Manual
Descriptions of the IEEE_FLOAT and
TANDEM_FLOAT pragmas
C/C++ Programmer’s Guide