Currently most RADIANT software identifiers use the FPGA's device DNA, which would be great if it was, well, actually unique, but the ID readback inside the FPGA sadly... isn't: it's only 57 bits of the 64-bit unique ID, and as I've recently learned, Xilinx is an ass and doesn't partition off the unique IDs intelligently, so collisions are actually more common than you'd expect.
The board manager, though, does have a fully-unique 128-bit identifier which can be accessed as follows:
Currently most RADIANT software identifiers use the FPGA's device DNA, which would be great if it was, well, actually unique, but the ID readback inside the FPGA sadly... isn't: it's only 57 bits of the 64-bit unique ID, and as I've recently learned, Xilinx is an ass and doesn't partition off the unique IDs intelligently, so collisions are actually more common than you'd expect.
The board manager, though, does have a fully-unique 128-bit identifier which can be accessed as follows:
https://microchip.my.site.com/s/article/Reading-unique-serial-number-on-SAM-D20-SAM-D21-SAM-R21-devices
It'd be pretty trivial to expose the ID registers at 0x30/0x34/0x38/0x3C (12/13/14/15 in the case list).