1.3.2. Processor Options

1.3.2.1. Select Processor

-mcpu=<arg>

Instruct the compiler to generate code for the C29x processor variant indicated by <arg>, where <arg> can be:

  • c29.c0

Since only one C29x option is currently recognized, this setting is the default. The -mcpu option is not required.

1.3.2.4. Select Floating-Point Code Generated

The C29x CPU can perform native 32-bit and 64-bit floating-point hardware operations on the CPU, without sending data to a separate FPU for processing.

-mfpu=<arg>

The C29x compiler toolchain emits native 32-bit floating point operations by default. Optionally, some C29x hardware supports native 64-bit floating point operations. Users may control the usage of native floating point hardware with the -mfpu option.

Valid forms of the -mfpu option include:

  • -mfpu=none - Emulate all floating point operations in software without the use of floating-point registers or hardware support.

  • -mfpu=f32 - Use native 32-bit floating-point hardware operations, but emulate 64-bit floating-point operations in software. This is the default setting for the compiler toolchain.

  • -mfpu=f64 - Use native 32-bit and 64-bit floating-point hardware operations.

1.3.2.5. Select Endianness

C29x devices are always little-endian. No command-line options are provide to specify or change this.