In Chapter 4 we discussed the differences between fixed-point and floating-point processors and the influence of hardware support for the computing time of numerical mathematics. In Lab 4 we also benchmarked the performance of the F2833x for fixed-point and floating-point implementations. The good thing with a Digital Signal Controller, such as the F2833x is that we can decide whether to generate fixed-point code or floating-point machine code, because the hardware of this device supports both worlds. There are not that many controllers in the market, which give us such flexibility! However, the C2000 family of Digital Signal Controller includes also some other members without the additional support of a floating-point unit. There is one other member, the TMS320F28035, which has an additional hardware "Control Law Accelerator" (CLA), which is also based on floating-point hardware. But for all other members of the C2000 - family, for example the F281x and the F280x groups, we cannot rely on a floating-point hardware.
The current version (version 1.5a) of Texas Instruments IQ-Math - Library can be found in literature number SPRC087 at www.ti.com.