Closing post doesn’t told us you are having this correct!
Please explain why it is wrong:
There are no requirement on composable overlay to involve PR block designs so DFX aren’t 1 of the requirement as it is just an AXI-Stream switch board on different processing units.
DFX is a requirement of composable? any PR block or P-block is required on the FIR example?
So from the above points it is clearly you are just telling people to silent wo clear out the concept here.
(Partial Reconfiguration — Python productivity for Zynq (Pynq))
Are there any relationship or topic on Mario link about “composable”?
The answer is NO
So you still fail to clear out does DFX API need AXI-Stream to work or not?
You cleanup the final part.
As you had mentioned DFX and composable overlay are completely two individual topic and APIs.
So I will expected top blocks are just individual and what the composable had done as you also mentioned are bring them together but does not necessary need DFX to work as plug board idea.
So why would AXI4-Stream become necessary when consider DFX API as your link do show no prerequisite of AXI4-Stream designs.
I am answering these question based on the context of your question, composable overlay.
If you want a composable overlay - given our definition of composable overlay - you need to use AXI4-Stream, and the python API only understand AXI4-Stream.
For general DFX, as said you can use any interface type.
Mario, what I understand is that the composable overlay DFX used the previous “General DFX” python API and build above it? (Correct me if this is not the case)
