Unable to run a circuit on real quantum processor

[gaoning041x]

Informations

• Qiskit version: 0.6.1
• Python version: 3.5.2
• Operating system: Ubuntu(64-bit)

What is the current behavior?

I have constructed a 5 qubit, 3 classical bit circuit. This circuit works fine if the backend is the simulator. But when I try to run it on the real quantum processor, it failed and print the following error message:

Traceback (most recent call last): File "quantum.py", line 89, in result_real = job_exp.result() File "/usr/local/lib/python3.5/dist-packages/qiskit/backends/ibmq/ibmqjob.py", line 195, in result job_response = self._wait_for_result(timeout=timeout, wait=wait) File "/usr/local/lib/python3.5/dist-packages/qiskit/backends/ibmq/ibmqjob.py", line 209, in _wait_for_result 'it is {}'.format(str(status))) qiskit.backends.joberror.JobError: 'Invalid job state. The job should be DONE but it is JobStatus.ERROR'

Steps to reproduce the problem

Just run the following program, then supposedly the error message will come out. By the way, I do have an IBMQ experience account with some free accesses to the real quantum processor.

import matplotlib.pyplot as plt import numpy as np import math from qiskit import QuantumCircuit, ClassicalRegister, QuantumRegister from qiskit import available_backends, execute, register, get_backend from qiskit.tools.visualization import plot_histogram, circuit_drawer from qiskit.tools.visualization import matplotlib_circuit_drawer as drawer, qx_color_scheme from qiskit.tools.qi.qi import state_fidelity from qiskit import Aer from qiskit import IBMQ

import the real machine

from qiskit.backends.ibmq import least_busy IBMQ.load_accounts() backend = Aer.get_backend("qasm_simulator")

q[0] is the control bit, q[1], q[2] are to be measured, q[3], q[4] are to construct quantum switch

on classical quantum computer.

q = QuantumRegister(5, "q") c0 = ClassicalRegister(1, "c0") c1 = ClassicalRegister(1, "c1") c2 = ClassicalRegister(1, "c2") qc = QuantumCircuit(q, c0, c1, c2, name = "qc")

prepare q[0] in superposition

qc.h(q[0])

put q[2] q[3] into \phi^+

qc.h(q[2]) qc.cx(q[2], q[3])

put q[1] into maximally mixed state

put q[4] into |+>

q[4] is only to generate a random grey code

qc.h(q[4]) qc.measure(q[4], c0[0]) qc.x(q[1]).c_if(c0, 1)

Controlled Swap q[1] q[2] conditioned on q[0]

qc.cswap(q[0], q[1], q[2])

flip q[0]

qc.x(q[0])

Measure q[1] in fourier basis

qc.h(q[1]) qc.measure(q[1], c0[0])

Measure q[2] in computational basis

qc.measure(q[2], c0[0])

Controlled Swap q[1] q[2] conditioned on q[0]

qc.cswap(q[0], q[1], q[2])

Measure q[2] q[3] on \phi^+

qc.cx(q[2], q[3]) qc.h(q[2])

qc.measure(q[2], c1[0]) qc.measure(q[3], c2[0])

measure q[0] in fourier basis

qc.h(q[0]) qc.measure(q[0], c0[0])

drawing

"""cmp_style = qx_color_scheme() cmp_style.update({ 'usepiformat': True, 'showindex': True, 'cregbundle': True, 'compress': True, 'fold': 20 }) drawer(qc, filename="AMDExample.pdf", style = cmp_style)"""

"""# Create a Quantum Program for execution job = execute(qc, backend, shots = 40000) print(job.status()) result = job.result() plot_histogram(result.get_counts('qc'))"""

running on a real machine

large_enough_devices = IBMQ.backends(filters=lambda x: x.configuration()['n_qubits'] > 5 and not x.configuration()['simulator']) backend = least_busy(large_enough_devices) print("The best backend is " + backend.name())

qiskit_job_status

shots = 8192 # Number of shots to run the program (experiment); maximum is 8192 shots. max_credits = 5 # Maximum number of credits to spend on executions.

job_exp = execute(qc, backend=backend, shots=shots, max_credits=max_credits) result_real = job_exp.result()

print(result_real.get_counts("qc"))

What is the expected behavior?

In expectation, the circuit should run successfully and the program should print a plot indicating the statistic.

Suggested solutions

I need help with this

[maddy-tod]

Hi @gaoning041x could you give some more information, which backend were you using? I don't think the physical devices currently support operations after measurements, so you may need to re-write your code or use one of the simulators instead.

[gaoning041x]

Hi @maddy-tod thank you for your reply. I was trying to run this circuit on the 16 qubit machine. I want to run this circuit on a 4 qubit machine, but I need an extra qubit to generate a qubit mixed state, like I/2. So do you mean that I cannot manipulate the state after I measure it? Can I just do something like measure the system on one basis, then measure it on another basis?

[maddy-tod]

Yes that is what I mean. I don't think that measuring like that is currently possible on the real backends. However I managed to get your code running using the Aer simulator, Aer.get_backend('qasm_simulator'), so it might be better to use this for now.

[gaoning041x]

@maddy-tod, thanks for clarifying. The circuit does work fine on simulators. Besides these problems, I would also like to ask one additional question if you don't mind: how can I generate a mixed state like I/2, how can I do this without introducing an extra qubit? I searched through the documentation and I cannot find it... Thanks for your help.

[rraymondhp]

Hi, IMHO, mixed states are not supported yet, but there are at least two workarounds: (1) By purification, you can work on corresponding pure states by doubling the number of qubits (2) Or, you can combine randomized coin flipping with the current pure states simulator/real-devices

[gaoning041x]

Hi @rraymondhp, thanks for the suggestions, but either one would require an extra qubit right? The circuit I'm using now essentially adopts the second suggestion: I use an extra qubit (to do coin tossing) and classical control to generate a mixed state. Anyway thanks for help.

[nonhermitian]

A mixed state is represented by a density matrix. On a real device you always have a state vector. A density matrix is what you get after performing infinitely many realizations of an experiment and averaging.

[qubeat]

In principle, conception of reduced density matrix (https://en.wikipedia.org/wiki/Quantum_entanglement#Reduced_density_matrices) may be applied even for single experiment due to entanglement with environment.

[rraymondhp]

Hi @gaoning041x The second one will not require extra qubits. And as @nonhermitian pointed out, density matrix is an average of many experiments on pure states.

So the conclusion is that, this issue is not a bug and can be closed, right?

[gaoning041x]

@rraymondhp Right, I don't need extra qubits in principle if I follow the second suggestion. The original problem I was asking is actually not about mixed states, but it's been answered by @maddy-tod and yes, I think my questions are all clear now and I'm happy with you closing this issue. Thanks.

[maddy-tod]

@gaoning041x great to hear! If you need any help in future the Qiskit Slack workspace is a great place to start.