Using: Arm Cortex-M3
IDE: Keil
When I placed code in ram, I could see the functions' executing address indeed is in ram address range from map file.
But interesting thing is I found the change of ram size used in map file's summarization is not an order of magnitude compares to the actual size of functions.
Why?
Working with little data here, if you want better answers present the question better.
Data copied to RAM will include subroutines and literal pool data. There will be a copy in FLASH that gets copied to RAM, likely via the _scatterload()
Show target memory organization. Show the relevant portions of .MAP, or all of it if you can't make that determination.
Show the areas you have issues with, and what your expectation for it would be.