rtel wrote on Tuesday, November 06, 2012:
I can’t be exactly sure where the libraries came from, but they will have either been included with the software package that came with the experimenter board used to host the demo, or from CCS4 itslef. Normally I would base a new project on a project provided by the board vendor, then strip out what I didn’t need - so that is the most likely case here. Have you asked on the TI forums?
I would recommend starting with the small data and code model as we are aware of issues when the FreeRTOS demo is used with CCS5 with a large data model. Once the small model demo is running you can move onto the large data and code models if that is necessary. I’m sure whatever is causing the problem will be easy to fix.
The reported problem appears to be related to the libraries. In CCS4 I added my own basic library functions because the ones that came with CCS4 were too slow and big. In CCS5 it seems the library I added myself inexplicably doesn’t work (it is just C code and runs in numerous other demos), and better success is found with the CCS5 library directly.
I remember in general CCS made very heavy work of running the exact same code as the identical IAR demo. I’m sure CCS5 has many improvements over CCS4 though.