Hardware floating point can be attractive in many embedded applications where integer or fixed point algorithms produce either overflow/underflow or unacceptably long word lengths. In such cases long design times are likely to be needed in order to ensure predictable accuracy and avoid scaling problems. The FPS26 from Cortus offers hardware single precision floating point that enables the treatment of signals with a wide dynamic range with the required precision and speed.
Cortus have analysed typical embedded programs and compiler output to ensure that your software is compiled with the very highest code density possible. In tests using standard code benchmarks the FPS26 shows an improvement of at least 16% over the FPS6.