Hello,
when I take an example from the SDK to run in RAM, the linker script results is created with the following section:
.text : ALIGN(4)
{
*(.text*)
*(.rodata .rodata.* .constdata .constdata.*)
. = ALIGN(4);
} > SRAM_ITC
I wonder if better performance would be achieved by moving .rodata* or .constdata* to DTC, since the instruction (to ITC) and data (to DTC) buses are separate.
What drawbacks should there be to doing this? And why are code and (readonly) data mixed in the SDK examples.
best reagrds
Max