High-level architecture question:

When we have the ability to do useful quantum computations with plenty of qubits, error correction and fault tolerance, do Clifford simulators still have a role to play? Can a quantum computer still delegate Clifford gate computations to a classical device and only perform those gates that cannot be efficiently classically simulated?

Or are these simulators just a way for us to test error correction and other Clifford-only operations and would not help reduce the resource requirements for an actual quantum computation?


1 Answer 1


There are potential uses for Clifford simulation even with a large scale quantum computer available. To help it run more smoothly.

For example, in Daniel Litinsky's "game of surface codes" paper, the quantum computer doesn't do the Clifford gates. Instead, the classical computer tracks a Clifford frame and rewrites T gates and measurement gates to apply to large Pauli products determined by the Cliffords. The Clifford tracking is essentially identical to stabilizer simulation.

Backdated Pauli corrections for delayed choice gates are also applied in the present by using stabilizer simulation to advance them through the intervening Clifford gates. This is core to AutoMagic states decoupling where the correction occurs from where the non-Clifford gate is applied.

Just understanding how errors flow through the circuit requires a bit of stabilizer simulation (though with only one Pauli string instead of many).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.