7.1.4 DO Loop Level

The switches, -inline_looplevel=<n> [-inll] and -ipa_looplevel=<n> [-ipall], set a minimum DO loop nest level for CALL/function reference expansion. The -inline_looplevel and -ipa_looplevel switches enable you to limit inlining and IPA to just routines that are referenced in nested loops, where the reduced call overhead or enhanced optimization will be multiplied.

The argument is defined from the most deeply nested leaf of the call tree. The default, 10, allows inlining (IPA) for the 10 deepest nest levels, for example:

 PROGRAM MAIN
  ..
   CALL A ---> SUBROUTINE A
 ..
    DO
     DO
       CALL B --> SUBROUTINE B
     ENDDO  DO
     ENDDO  DO
        CALL C ---> SUBROUTINE C
     ENDDO
    ENDDO

The CALL B is inside a doubly nested loop, and would be more profitable to expand than the CALL A . The CALL C is quadruply nested, so inlining C would yield the biggest gain of the three.

The argument is defined from the most deeply nested CALL or function reference:

• -inline_looplevel=1 says only the routines referenced in the most deeply nested call site(s) can be inlined, for example, SUBROUTINE C . If more than one routine is called at the same loop nest level, all of them are selected when that level is included.

• -inline_looplevel=2 says only routines CALLED at the most deeply nested level and one loop less deeply nested can be inlined. The -inline_looplevel=3 setting would be required to inline SUBROUTINE B , since its call is two loops less nested than the call to SUBROUTINE C . A value of 3 or greater will cause C to be inlined into B , then the new B to be inlined into the main program.

• Setting -inline_looplevel (or -ipa_ looplevel ) to a large value will permit inlining at any nesting level.

• The calling tree written to the listing file with -listoptions=c includes the nesting depth level of each call in each program unit and the aggregate nesting depth which is the sum of the nesting depths for each call site, starting from the main program. This information can be used to identify the best routines for inlining.

• A routine that passes the -inline_looplevel test will be inlined everywhere it is used, even in places that are not in deeply nested loops. If some, but not all, invocations of a routine are to be expanded, use the inline and IPA directives just before each CALL/reference that is to be expanded. (See Section 7.1.5.)

• Because inlining increases the size of the code, the extra paging and cache contention can actually slow down a program. Restricting inlining to routines used in DO loops multiplies the benefits of eliminating subroutine and function call overhead for a given amount of code space expansion. If inlining appears to have slowed an application code, investigate using IPA, which has little effect on code space and the number of temporary variables.