AnsweredAssumed Answered

Floating-point divide needs a library function?

Question asked by rallen911 on Sep 10, 2010
Latest reply on Sep 13, 2010 by CraigG

We are using the ADSP-21369 processor for our next design, and we are reusing some software that was written for the ADSP-21161 using v3.5 of Visual DSP++.  In that software, there is a floating point divide.  When I try to include this file in our build, the v5.0 compiler wants to include the library call __lib_divf.  I went back to the old project, and that library was not included in the software.


Normally, I wouldn't really think too much about it, but our products are safety critical avionics applications and require complete requirements and design documentation for all software functions contained in the application.  We would like to avoid this work if the library doesn't truly need to be included.


Why was it not necessary in the previous project?  Is it processor dependent?  Is it a diffence in the tools?




Robert Allen

Goodrich Corp