Forum Discussion

dj-park's avatar
dj-park
Icon for Occasional Contributor rankOccasional Contributor
3 years ago
Solved

What's special about Static Update Partial Reconfiguration?

I am comfortable with AMD(Xilinx) Partial Reconfiguration but new to Intel PR. I've found "Static Update Partial Reconfiguration" feature for Intel PR. But I don't understand how different it is to...
  • sstrell's avatar
    sstrell
    3 years ago

    Again, if you think you might want or need to make changes to the static region, which normally must be completely fixed and not changed at all, using SUPR can save you some time and headache. Also note that some resources in the static region, like I/O, cannot be part of a PR region so they can't be PR'ed during runtime anyway.

    If something in your static region can be part of a PR region, then sure, create another PR region for it, isolating it from the static region.

sstrell's avatar
sstrell
Icon for Super Contributor rankSuper Contributor
3 years ago

The idea behind SUPR is that normally, if you change anything in the static region of a PR design, you have to completely recompile the static and PR regions (base and implementation revisions). With SUPR, you can change part of the static region without having to do all the recompiling. The difference from just creating another PR region is that the SUPR region cannot be switched to other logic during runtime. Standard PR regions can be swapped with other personas at any time while the design is running.

The main idea is to save compiles if you think you may need to make changes to your static region design. See the app note: https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/an/an818.pdf