This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

maptool.py versus RTSC for multicore deployment

Is there a preferred idiom or approach for deploying multicore applications with RTSC?  Right now it looks like the simplest approach is to build the individual executables (libraries and applications) using RTSC, then separately use maptool.py to combine the executables into the combined image.  Would it make sense to create an RTSC xdc.bld.ITarget or xdc.bld.ITargetFilter provider to expose maptool.py's functionality so the development process is more integrated?

  • What is maptool.py? Is there any documentation we could look at?

  • maptool.py is a Python script that takes a number of ELF images, plus directions on how they should be mapped to the cores in a multi-core device (and how to set up memory mappings for each), and generates a combined image that the MCSDK "intermediate bootloader" (IBL) will load and launch.  There is documentation on maptool.py and the rest of that framework at http://processors.wiki.ti.com/index.php/MAD_Utils_User_Guide .

    As I understand it, maptool.py does two main things: First, it (loosely speaking) "relinks" the applications so that their data items in global memory do not overlap, while local data is put on the correct core (for example, App1 running on core 1 and App2 running on core 2 can use the same L2 SRAM address -- in their respective cores -- for different variables).  Second, it generates a register load table that the IBL and associated code use to configure the MPAX (memory protection), start address, and so on for each core.