Implementing the bridge gate.

[Djakam]

Added a new gate which is the bridge gate, it is named brg

Summary

Details and comments

[CLAassistant]

All committers have signed the CLA.

[kdk]

Thanks @Djakam , this looks great so far! A few minor comments below.

[kdk]

If you know the __array__ representation, it would be good to add this as a case to TestGateDefinitions in test/python/circuit/test_gate_definitions.py (and even if not, it can be added to TestStandardGates.test_definition_parameters in the same file to ensure the definition and equivalence library entires remain in sync).

[Djakam]

Thanks, @kdk. I will change the code accordingly before pushing it again.

[kdk]

Thanks for the updates here @Djakam , this looks great! The lass piece will be to expand the gate name from brg to bridge (and likewise for the class name).

[ajavadia]

should probably expand this to long-range CNOT on more than 3 qubits. e.g. for 4 qubits it would look like this:

[Djakam]

@ajavadia , thanks for your input. What could be the name for the 4 qubits? what about the 5 ? Any propositions?

[ajavadia]

I'm just trying to make this a bit more general than this very specific use case. I'm certainly not suggesting making a 3, 4, 5, qubit version of this gate.

So the Bridge gate implemented thus far in your PR is a synthesis of a CNOT on 2 qubits separated by one qubit, on a line. The more general thing is you want to synthesize a CNOT on 2 qubits that are distant on some coupling_map. I think there are algorithms that can do this efficiently for the general case, and give you the bridge gate for this specific case.

On the other hand if you want to do the circuit optimization suggested in #6656, then I think a better approach is to use the template optimizations in Qiskit. Just define a template and the template optimization pass will do it, no need to write a separate pass.

[kevinsung]

Why does this inherit from QuantumCircuit instead of Gate?