Ensemble average in quantum computing

[morphemera]

In the last step to get result from a quantum computer, measurement is required to collapse the quantum state. This can be done by the measuring a large ensemble average of same computed result. One realization I know is to use NMR to measure billions of spin. So, what is the minimum number of ensembles are required to get a 'good' result?

Furthermore, suppose the number of one qubit requires O(1) ensemble, how many ensembles are required for n qubits system to have same accuracy? I suppose the answer should be O(n), but if it is O(2^n), then the realization of quantum computer still seems useless to me. Any explanation of the answer?

 

[sokrates]

Is there a minimum number as such?

Let me ask you this: What's a good enough number of coin flips to see almost half the flips are heads?

You are talking about an accuracy in your second question, how much accuracy are you aiming for?

My point is there's no fundamental minimum. It's exactly like a coin toss experiment, nothing peculiar to quantum computing here, in this context.