Instruction Scheduling

You can schedule instructions optimally for specific processor instruction latencies and cache sizes. Using any of the options in the following table produces binary code that will run on all processors, but that is optimized for the target processor(s).

Option Setting

Processor

Windows*

Linux*

-G5

-tpp5

Pentium® processors and Pentium processors with MMX™ technology

-G6

-tpp6

Pentium Pro, Pentium II and Pentium III processors

-G7 default

-tpp7 default

Pentium 4 processor

When using these options, the compiler optimizes by executing the required math operation with the least costly operation, based on the shift/multiply latency of the target processor.