ColdFire Compiler > Optimization

Use this panel to control compiler optimizations. The compiler's optimizer can apply any of its optimizations in either global or non-global optimization mode. You can apply global optimization at the end of the development cycle, after compiling and optimizing all source files individually or in groups.

The following table lists and defines each option of the Optimization panel.

Table 1. Tool Settings - ColdFire Compiler > Optimization Options
Option Description
Optimization Level (-opt) Specify the optimizations that you want the compiler to apply to the generated object code:
  • 0-Disable optimizations. This setting is equivalent to specifying the -O0 command-line option. The compiler generates unoptimized, linear assembly-language code.
  • 1-The compiler performs all target-independent (that is, non-parallelized) optimizations, such as function inlining. This setting is equivalent to specifying the -O1 command-line option.

    The compiler omits all target-specific optimizations and generates linear assembly-language code.

  • 2-The compiler performs all optimizations (both target-independent and target-specific). This setting is equivalent to specifying the -O2 command-line option. The compiler outputs optimized, non-linear, parallelized assembly-language code.
  • 3-The compiler performs all the level 2 optimizations, then the low-level optimizer performs global-algorithm register allocation. This setting is equivalent to specifying the -O3 command-line option. At this optimization level, the compiler generates code that is usually faster than the code generated from level 2 optimizations.
Speed Vs Size

Use to specify an Optimization Level greater than 0.

  • Speed-The compiler optimizes object code at the specified Optimization Level such that the resulting binary file has a faster execution speed, as opposed to a smaller executable code size.
  • Size-The compiler optimizes object code at the specified Optimization Level such that the resulting binary file has a smaller executable code size, as opposed to a faster execution speed. This setting is equivalent to specifying the -Os command-line option.
IPA Specifies the Interprocedural Analysis (IPA) policy.
  • Off - No interprocedural analysis, but still performs function-level optimization. Equivalent to the "no deferred inlining" compilation policy of older compilers.
  • File - Completely parse each translation unit before generating any code or data. Equivalent to the "deferred inlining" option of older compilers. Also performs an early dead code and dead data analysis in this mode. Objects with unreferenced internal linkages will be dead-stripped in the compiler rather than in the linker.
Inlining Enables inline expansion. If there is a #pragma INLINE before a function definition, all calls of this function are replaced by the code of this function, if possible. Using the -Oi=c0 option switches off inlining. Functions marked with the #pragma INLINE are still inlined. To disable inlining, use the -Oi=OFF option.
Bottom-up Inlining Check to control the bottom-up function inlining method. When active, the compiler inlines function code starting with the last function in the chain of functions calls, to the first one.