You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The high fanout net routing code currently counts both the total nodes_added and chan_nodes_added in different tests to see if we need to add more of the partial routing. This looks wrong: we should always be using chan_nodes_added for all tests and the nodes_added variable should be deleted.
I think this code is messing up the parallel router with net decomposition and flat routing (although that has not been confirmed yet). In any case, this change would simplify the code slightly and make it more logical.
Expected Behaviour
Make this change, test QoR is OK. @duck2 should then check flat routing with net decomposition again.
Context
It has potentially caused @duck2 mental anguish :).
The text was updated successfully, but these errors were encountered:
I think it makes sense to use chan_nodes_added in the tests (line 1057) and delete nodes_added variable.
It appears (line 1033) that only the flat router checks if the node is a wire before incrementing chan_nodes_added. Unless I am missing something, I would expect this check to be performed irrespective of whether the router is flat or not. What do you think?
The high fanout net routing code currently counts both the total nodes_added and chan_nodes_added in different tests to see if we need to add more of the partial routing. This looks wrong: we should always be using chan_nodes_added for all tests and the nodes_added variable should be deleted.
I think this code is messing up the parallel router with net decomposition and flat routing (although that has not been confirmed yet). In any case, this change would simplify the code slightly and make it more logical.
Expected Behaviour
Make this change, test QoR is OK.
@duck2 should then check flat routing with net decomposition again.
Context
It has potentially caused @duck2 mental anguish :).
The text was updated successfully, but these errors were encountered: