l
ld(1) | ld(1) |
O (o) | Trap on |
U (u) | Trap on |
I (i) | Trap on |
D (d) | Enable sudden under¯ow (¯ush to zero) of denormalized values. |
| Note: Enabling sudden under¯ow is an unde®ned operation on |
| |
| Selecting this ¯ag enables sudden under¯ow only if it is available on the |
| processor being used at |
To dynamically change these settings at
+I symbol Specify the name of the initializer function when building a shared library. A shared library may have multiple initializers speci®ed. Initializers are executed in the order that they are speci®ed. You can specify more than one initializer, but each must be preceded by +I. For more details on the initializer function, refer to the +help option or the
+O[no]fastaccess
Enable [disable] fast access to global data. The linker rearranges the data to further reduce the number of ADDIL instructions in the executable.
This optimization may reveal some subtle programming errors related to assumptions about data layout. This optimization can occur at optimization levels 2, 3 and 4. The default is +Onofastaccess at optimization levels 2 and 3, and +Ofastaccess at optimization level 4.
Using the +Ofastaccess option in the presence of debug information generates a warning message and can cause corruption of the debug information. Avoid using +Ofastaccess with the
This option is accepted but ignored by the
+O[no]procelim
Enable [disable] the elimination of procedures that are not referenced by the applica- tion. The default is +Onoprocelim . Procedure elimination can occur at any optimization level, including level 0. For more details refer to the +help option or the
+Oreusedir= dir
Specify the name of the directory to use as the reuse repository for the object code reuse feature. This directory holds the compiled object ®les for reuse. dir can be an absolute or relative path to directory where you invoked the linker. The +Oreusedir option shortens link time by not recompiling intermediate object code to object code when using +O4 or
+Oselectivepercent n
Instructs the interprocedural optimizer driver to pass the ®rst n percent of the object ®les to the high level optimizer for interprocedural optimizations such as inlining.
+Oselectivesize size
Instructs the interprocedural optimizer driver to pass the ®rst k routines to the high level optimizer for interprocedural optimization where the size of k routines are approaching but less than size.
+OselectiveO3
Instructs the interprocedural optimizer driver to compile the routines not included in the +O4 list to be compiled at +O3.
+Ostaticprediction
Meaningful only on PA 2.0 architecture, this option sets the branch prediction bit in the output executable ®le's auxiliary header.
32 Bit Link Editor Options
Section 1−430 | − 8 − |