Making RQC generator output compatible with qFlex input.

[s-mandra]

No description provided.

[s-mandra]

See #101.

[alexandrupaler]

Is this still an issue?

[s-mandra]

It is still an issue. I have few workaround that could be potentially implemented, but it would be better if the rqc generator would output the circuit using the same indexes for qubits that qflex actually uses.

[viathor]

What is the current qubit indexing scheme in qflex? I recall there had been plans to change it...

[s-mandra]

qflex uses the provided grid to number the qubits. For instance, for the IBM rochester, the grid can be found in config/grid and numbers are assigned to qubits regardless if they are active or not in the grid (starting from the top left to the bottom right). The RQC assigns instead numbers only for the active qubits. To this end, the RQC could accept a grid as an input and then reindex everything accordingly to the grid.

[viathor]

I have a recollection of us deciding (or at least considering) to use contiguous indices for consistency with qsim. Do we still intend to do that?

If so, then instead of changing the RQC generator, I can change qflex to use contiguous indices in the input files.

[s-mandra]

I have a recollection of us deciding (or at least considering) to use contiguous indices for consistency with qsim. Do we still intend to do that?

If so, then instead of changing the RQC generator, I can change qflex to use contiguous indices in the input files.

At the moment, qflex is compatible with cirq (which doesn’t use contiguous indexes) so I would keep the compatibility with cirq rather than with qsim. In my opinion, I would prefer to add a flag to the rqc to either output the circuit using contiguous indexes or following a specific grid.