The DSP56800E extends the DSP56800 data addressing range, by providing 24-bit address capability to some instructions. 24-bit address modes allow user accesses beyond the 64K-word boundary of 16-bit addressing. To control large data memory model support, use the M56800E Processor panel as shown in the following figure. See DSC Compiler > Processor for explanations of this panel’s elements.
Extended data is data located beyond the 16-bit address boundary — as if it exists in extended (upper) memory. Memory located below the 64K boundary is lower memory.
The compiler default arrangement is using 16-bit addresses for all data accesses. This means that absolute addresses (X:xxxx addressing mode) are limited to 16 bits. Direct addressing or pointer registers load or store 16-bit addresses. Indexed addressing indexes are 16-bit quantities. The compiler treats data pointers as 16-bit pointers that you may store in single words of memory.
Linker support is present to error out with correct help message for data access above 0x7FFF (without -largeAddrInSdm option usage) for small data model cases.
If the large data memory model is enabled, the compiler accesses all data by 24-bit addressing modes. It treats data pointers as 24-bit quantities that you may store in two words of memory. Absolute addressing occurs as 24-bit absolute addresses. Thus, you may access the entire 24-bit data memory, locating data objects anywhere.
You do not need to change C source code to take advantage of the large data memory model.
Examples in DSP56800E assembly code of extended data addressing are: