Big Time

 Really hitting the Big Time with this tutorial code on error mitigation. I am asking

for 100,000 shots and getting the answer back in minutes. 

-The circuit is parametrizedfor efficiency, alternating single gates with 2-qubit gates.

(The values on the gates are random...).

-A mirror circuit is constructed. Without a barrier, the transpiler would create a circuit

with no gates; this setup runs the identity matrix.

-The transpiler does the work of optimizing the circuit, using 'generate_preset_pass_manager'. 

Each 'pass' looks for a given loose end it can tidy up, or simpify in the code. Of course, one might

design one's own pass manager.

-The whole is run from the Estimator primitive.

-One uses ISA configuration.

The code is really 5 jobs run in 'batch' mode. One with no error mitigation, then one

adds each error mitigation technique up to maximum. 

I am used to running code that takes a few seconds. I ran this code twice, one for 10

qubits and one for 50, and each took over 4 minutes to run. I have completely used

up my monthly allocation of computer time and can only get back on December1...

But that is really the whole point of the exercise. One can ask for more and better error

mitigation, but it will create overhead costs because it takes more computer time to run.

Below:

                                                                           

                                                                                

Result for a 50 qubit circuit:

                                                                                   

-I seem to have escaped error bars at the top of my results!?

Combine error mitigation options with the estimator primitive | IBM Quantum Learning