IQM Resonance#

1. Login to IQM Resonance with your credentials.

2. Upon your first visit to IQM Resonance, you can generate your unique, non-recoverable API token directly from the Dashboard page by selecting Generate token. It’s important to copy the token immediately from the window, as you won’t be able to do so once the window is closed. If you lose your token, you have the option to regenerate it at any time. However, be aware that regenerating your API token will invalidate any previously generated token.

3. Download one of the demo notebooks from IQM Academy or the resonance_example.py example file (Save Page As…)

4. Install Qiskit on IQM as instructed below.

5. Add your API token to the example (either as the parameter token to the IQMProvider constructor, or by setting the environment variable IQM_TOKEN)

6. Run the Jupyter notebook (or run python resonance_example.py if you decided to go for the Python script).

7. If you’re connecting to a real quantum computer, the output should show almost half of the measurements resulting in ‘00000’ and almost half in ‘11111’ - if this is the case, things are set up correctly!

You can find a video guide on how to set things up here. More ready-to-run examples can also be found at IQM Academy.

On-premises device#

1. Download the bell_measure.py example file (Save Page As…).

2. Install Qiskit on IQM as instructed below.

3. Install Cortex CLI and log in as instructed in the documentation

4. Set the environment variable as instructed by Cortex CLI after logging in.

5. Run $ python bell_measure.py --cortex_server_url https://demo.qc.iqm.fi/cocos - replace the example URL with the correct one.

6. If you’re connecting to a real quantum computer, the output should show almost half of the measurements resulting in ‘00’ and almost half in ‘11’ - if this is the case, things are set up correctly!

IQM Resonance#

If you are using IQM Resonance, you have two options to authenticate:

1. Set the IQM_TOKEN environment variable to the API token obtained from the Resonance dashboard.

2. Pass the token parameter to IQMProvider. This will be forwarded to IQMClient. For an example, see the resonance_example.py file

On-premises devices#

If the IQM server you are connecting to requires authentication, you may use Cortex CLI to retrieve and automatically refresh access tokens, then set the IQM_TOKENS_FILE environment variable, as instructed, to point to the tokens file. See Cortex CLI’s documentation for details.

You may also authenticate yourself using the IQM_AUTH_SERVER, IQM_AUTH_USERNAME and IQM_AUTH_PASSWORD environment variables, or pass them as arguments to IQMProvider, however this approach is less secure and considered deprecated.

Executing a circuit#

Let’s consider the following quantum circuit which prepares and measures a GHZ state:

from qiskit import QuantumCircuit

circuit = QuantumCircuit(3)
circuit.h(0)
circuit.cx(0, 1)
circuit.cx(0, 2)
circuit.measure_all()

print(circuit.draw(output='text'))

        ┌───┐           ░ ┌─┐
   q_0: ┤ H ├──■────■───░─┤M├──────
        └───┘┌─┴─┐  │   ░ └╥┘┌─┐
   q_1: ─────┤ X ├──┼───░──╫─┤M├───
             └───┘┌─┴─┐ ░  ║ └╥┘┌─┐
   q_2: ──────────┤ X ├─░──╫──╫─┤M├
                  └───┘ ░  ║  ║ └╥┘
meas: 3/═══════════════════╩══╩══╩═
                           0  1  2

To run this circuit on an IQM quantum computer you need to initialize an IQMProvider instance with the IQM server URL, use it to retrieve an IQMBackend instance representing the quantum computer, and use Qiskit’s transpile() function followed by IQMBackend.run() as usual. shots denotes the number of times the quantum circuit(s) are sampled:

from qiskit import transpile
from iqm.qiskit_iqm import IQMProvider

iqm_server_url = "https://demo.qc.iqm.fi/cocos/"  # Replace this with the correct URL
provider = IQMProvider(iqm_server_url)
backend = provider.get_backend()

transpiled_circuit = transpile(circuit, backend=backend)
job = backend.run(transpiled_circuit, shots=1000)

You can optionally provide IQMBackend specific options as additional keyword arguments to IQMBackend.run(), documented at IQMBackend.create_run_request(). For example, you can enable heralding measurements using circuit_compilation_options as follows:

from iqm.iqm_client import CircuitCompilationOptions

job = backend.run(transpiled_circuit, shots=1000, circuit_compilation_options=CircuitCompilationOptions(heralding_mode=HeraldingMode.ZEROS))

Calibration#

The calibration data for an IQM quantum computer is stored in a calibration set. An IQMBackend instance always corresponds to a specific calibration set, so that its transpilation target uses only those QPU components (qubits and computational resonators) and gates which are available in that calibration set. The server default calibration set will be used by default, but you can also use a different calibration set by specifying the calibration_set_id parameter to IQMProvider.get_backend() or IQMBackend. If the server default calibration set has changed after you have created the backend, the backend will still use the original default calibration set when submitting circuits for execution. When this happens you will get a warning. You will need to create a new backend if you want to use the new default calibration set instead.

Inspecting the results#

The results of a job that was executed on the IQM quantum computer, represented as a Result instance, can be inspected using the usual Qiskit methods:

result = job.result()
print(result.get_counts())
print(result.get_memory())

The result comes with some metadata, such as the RunRequest that produced it in result.request. The request contains e.g. the qubit mapping and the ID of the calibration set that were used in the execution:

print(result.request.qubit_mapping)
print(result.request.calibration_set_id)

[
  SingleQubitMapping(logical_name='0', physical_name='QB1'),
  SingleQubitMapping(logical_name='1', physical_name='QB2'),
  SingleQubitMapping(logical_name='2', physical_name='QB3')
]
1320eae6-f4e2-424d-b299-ef82d556d2c3

Another piece of useful metadata are the timestamps of the various steps of processing the job. The timestamps are stored in the dict result.timestamps. The job processing has three steps,

• compile where the circuits are converted to instruction schedules,

• submit where the instruction schedules are submitted for execution, and

• execution where the instruction schedules are executed and the measurement results are returned.

The dict contains a timestamp for the start and end of each step. For example, the timestamp of starting the circuit compilation is stored with key compile_start. In the same way the other steps have their own timestamps with keys consisting of the step name and a _start or _end suffix. In addition to processing step timestamps, there are also timestamps for the job itself, job_start for when the job request was received by the server and job_end for when the job processing was finished.

If the processing of the job is terminated before it is complete, for example due to an error, the timestamps of processing steps that were not taken are not present in the dict.

For example:

print(result.timestamps['job_start'])
print(result.timestamps['compile_start'])
print(result.timestamps['execution_end'])

Backend properties#

The IQMBackend instance we created above provides all the standard backend functionality that one expects from a backend in Qiskit. For this example, I am connected to an IQMBackend that features a 5-qubit chip with star-like connectivity:

      QB1
       |
QB2 - QB3 - QB4
       |
      QB5

Let’s examine its basis gates and the coupling map through the backend instance

print(f'Native operations of the backend: {backend.operation_names}')
print(f'Coupling map of the backend: {backend.coupling_map}')

Native operations of the backend: ['id', 'r', 'cz', 'measure']
Coupling map of the backend: [[0, 2], [2, 0], [1, 2], [2, 1], [2, 3], [3, 2], [2, 4], [4, 2]]

Note that for IQMBackends the identity gate id is not actually a gate that is executed on the device and is simply omitted. At IQM we identify qubits by their names, e.g. ‘QB1’, ‘QB2’, etc. as demonstrated above. In Qiskit, qubits are identified by their indices in the quantum register, as you can see from the printed coupling map above. Most of the time you do not need to deal with IQM-style qubit names when using Qiskit, however when you need, the methods IQMBackendBase.qubit_name_to_index() and IQMBackendBase.index_to_qubit_name() can become handy.

Classically controlled gates#

Some IQM quantum computers support classically controlled gates, that is, gates that are executed conditionally depending on the result of a measurement preceding them in the quantum circuit. This support currently has several limitations:

• Only the x, y, rx, ry and r gates can be classically controlled.

• The gates can only be conditioned on one classical bit, and the only control available is to apply the gate if the bit is 1, and apply an identity gate if the bit is 0.

• The availability of the controlled gates depends on the instrumentation of the quantum computer.

The classical control can be applied on a circuit instruction using c_if():

from qiskit import QuantumCircuit

qr = QuantumRegister(2, 'q')
cr = ClassicalRegister(1, 'c')
circuit = QuantumCircuit(qr, cr)

circuit.h(0)
circuit.measure(0, cr[0])
circuit.x(1).c_if(cr, 1)
circuit.measure_all()

print(circuit.draw(output='text'))

        ┌───┐┌─┐        ░ ┌─┐
   q_0: ┤ H ├┤M├────────░─┤M├───
        └───┘└╥┘ ┌───┐  ░ └╥┘┌─┐
   q_1: ──────╫──┤ X ├──░──╫─┤M├
              ║  └─╥─┘  ░  ║ └╥┘
              ║ ┌──╨──┐    ║  ║
   c: 1/══════╩═╡ 0x1 ╞════╬══╬═
              0 └─────┘    ║  ║
meas: 2/═══════════════════╩══╩═
                           0  1

The first measurement operation stores its result in the 1-bit classical register c. If the result is 1, the X gate will be applied. If it is zero, an identity gate of corresponding duration is applied instead.

Executing the above circuit should result in the counts being approximately 50/50 split between the ‘00 0’ and ‘11 1’ bins of the histogram (even though the state itself is never entangled).

Resetting qubits#

The qiskit.circuit.Reset operation can be used to reset qubits to the $|0⟩$ state. It is currently implemented as a (projective) measurement followed by a classically controlled X gate conditioned on the result, and is only available if the quantum computer supports classically controlled gates.

from qiskit import QuantumCircuit

circuit = QuantumCircuit(1, 1)
circuit.h(0)
circuit.reset(0)
circuit.measure(0, 0)

print(circuit.draw(output='text'))

     ┌───┐     ┌─┐
  q: ┤ H ├─|0>─┤M├
     └───┘     └╥┘
c: 1/═══════════╩═
                0

In the above example, the Hadamard gate prepares a uniform superposition of the $|0⟩$ and $|1⟩$ states, and the reset then collapses it back into the $|0⟩$ state. Executing the circuit should result in (mostly) zeros being measured.

Inspecting circuits before submitting them for execution#

It is possible to inspect the final circuits that would be submitted for execution before actually submitting them, which can be useful for debugging purposes. This can be done using IQMBackend.create_run_request(), which returns a RunRequest containing the circuits and other data. The method accepts the same parameters as IQMBackend.run().

# inspect the run_request without submitting it for execution
run_request = backend.create_run_request(transpiled_circuit, shots=10)
print(run_request)

# the following two calls submit exactly the same run request for execution on the server
backend.run(transpiled_circuit, shots=10)
backend.client.submit_run_request(run_request)

It is also possible to print a run request when it is actually submitted by setting the environment variable IQM_CLIENT_DEBUG=1.

Basic transpilation#

You can use the default Qiskit transpiler on IQM quantum computers with both the Crystal and the Star architectures. Starting from the GHZ circuit we created above:

from qiskit.compiler import transpile

transpiled_circuit = transpile(circuit, backend=backend, layout_method='sabre', optimization_level=3)
print(transpiled_circuit.draw(output='text', idle_wires=False))

global phase: 3π/2
          ┌─────────────┐                  ┌─────────────┐ ░       ┌─┐
q_2 -> 5  ┤ R(π/2,3π/2) ├──────────■───────┤ R(π/2,5π/2) ├─░───────┤M├
          ├─────────────┤          │       └─────────────┘ ░ ┌─┐   └╥┘
q_0 -> 10 ┤ R(π/2,3π/2) ├─■────────■───────────────────────░─┤M├────╫─
          ├─────────────┤ │ ┌─────────────┐                ░ └╥┘┌─┐ ║
q_1 -> 15 ┤ R(π/2,3π/2) ├─■─┤ R(π/2,5π/2) ├────────────────░──╫─┤M├─╫─
          └─────────────┘   └─────────────┘                ░  ║ └╥┘ ║
  meas: 3/════════════════════════════════════════════════════╩══╩══╩═
                                                            0  1  2

Under the hood the Qiskit transpiler uses the IQMDefaultSchedulingPlugin plugin that automatically adapts the transpiled circuit to the IQMBackend. In particular,

• if optimization_level > 0, the plugin will use the IQMOptimizeSingleQubitGates pass to optimize single-qubit gates, and

• for devices that have the IQM Star architecture, the plugin will use the IQMNaiveResonatorMoving pass to automatically insert MoveGate instructions as needed.

Alternatively, you can use the transpile_to_IQM() function for more precise control over the transpilation process as documented below.

It is also possible to use one of our other pre-defined transpiler plugins as an argument to transpile(), for example transpile(circuit, backend=backend, scheduling_method="only_move_routing_keep"). Additionally, you can use any of our transpiler passes to define your own qiskit.transpiler.PassManager if you want to assemble custom transpilation procedures manually.

Computational resonators#

The IQM Star architecture includes computational resonators as additional QPU components, and uses qubit-resonator gates instead of two-qubit gates. These include MoveGate which moves qubit states to and from the resonators.

The standard Qiskit transpiler does not know how to compile qubit-resonator gates. This is why IQMBackend provides the Qiskit transpiler a simplified transpilation target in which the resonators and MOVE gates have been abstracted away, and replaced with fictional two-qubit gates that directly connect qubits that can be made to interact via a resonator. It then uses IQMDefaultSchedulingPlugin to re-introduce resonators and add MOVE gates between qubits and resonators as necessary at the scheduling stage.

IQMDefaultSchedulingPlugin is executed automatically when you use the Qiskit transpiler. Starting from the GHZ circuit we created above:

from qiskit.compiler import transpile
from iqm.qiskit_iqm import IQMProvider

resonator_backend = IQMProvider("https://cocos.resonance.meetiqm.com/deneb").get_backend()
transpiled_circuit = transpile(circuit, resonator_backend)

print(transpiled_circuit.draw(output='text', idle_wires=False))

           ┌─────────────┐┌───────┐                  ┌───────┐                ░ ┌─┐
  q_0 -> 0 ┤ R(π/2,3π/2) ├┤0      ├──────────────────┤0      ├────────────────░─┤M├──────
           ├─────────────┤│       │   ┌─────────────┐│       │                ░ └╥┘┌─┐
  q_1 -> 1 ┤ R(π/2,3π/2) ├┤       ├─■─┤ R(π/2,5π/2) ├┤       ├────────────────░──╫─┤M├───
           ├─────────────┤│  Move │ │ └─────────────┘│  Move │┌─────────────┐ ░  ║ └╥┘┌─┐
  q_2 -> 2 ┤ R(π/2,3π/2) ├┤       ├─┼────────■───────┤       ├┤ R(π/2,5π/2) ├─░──╫──╫─┤M├
           └─────────────┘│       │ │        │       │       │└─────────────┘ ░  ║  ║ └╥┘
resonators ───────────────┤1      ├─■────────■───────┤1      ├───────────────────╫──╫──╫─
                          └───────┘                  └───────┘                   ║  ║  ║
   meas: 3/══════════════════════════════════════════════════════════════════════╩══╩══╩═
                                                                             0  1  2

Custom transpilation#

As an alternative to the native Qiskit transpiler integration, you can use the transpile_to_IQM() function. It is meant for users who want at least one of the following:

• more fine grained control over the transpilation process without having to figure out which IQM transpiler plugin to use,

• transpile Star architecture circuits that already contain qubit-resonator gates, or

• force the transpiler to use a strict subset of qubits on the device.

For example, if you want to transpile the circuit with optimization_level=0 but also apply the single qubit gate optimization pass, you can do one of the following, equivalent things:

transpile_to_IQM(circuit, backend=backend, optimization_level=0, perform_move_routing=False, optimize_single_qubits=True)

transpile(circuit, backend=backend, optimization_level=0, scheduling_method='only_rz_optimization')

Similarly, if you want to transpile a native Star architecture circuit that already contains MoveGate instances (that act on a qubit and a computational resonator), you can do the following:

from iqm.iqm_client.transpile import ExistingMoveHandlingOptions
from iqm.qiskit_iqm import IQMCircuit, transpile_to_IQM

move_circuit = IQMCircuit(3)
move_circuit.h(0)
move_circuit.move(0, 1)
move_circuit.h(2)
move_circuit.cz(2, 1)
move_circuit.h(2)
move_circuit.move(0, 1)

# Using transpile() does not work here, as the circuit already contains a MoveGate
transpiled_circuit = transpile_to_IQM(move_circuit, backend=resonator_backend, existing_moves_handling=ExistingMoveHandlingOptions.KEEP)
print(transpiled_circuit.draw(output='text', idle_wires=False))

         ┌─────────────┐┌───────┐   ┌───────┐
q_0 -> 0 ┤ R(π/2,3π/2) ├┤0      ├───┤0      ├───────────────
         ├─────────────┤│       │   │       │┌─────────────┐
q_2 -> 1 ┤ R(π/2,3π/2) ├┤  Move ├─■─┤  Move ├┤ R(π/2,5π/2) ├
         └─────────────┘│       │ │ │       │└─────────────┘
q_1 -> 6 ───────────────┤1      ├─■─┤1      ├───────────────
                        └───────┘   └───────┘

And if you want force the compiler to use a strict subset of qubits on the device, you can do the following:

qubits = [4, 3, 8]
# or qubits = ['QB5', 'QB4', 'QB9']
transpiled_circuit = transpile_to_IQM(circuit, backend=backend, restrict_to_qubits=qubits)
print(transpiled_circuit.draw(output='text', idle_wires=False))

global phase: 3π/2
         ┌─────────────┐   ┌─────────────┐                ░    ┌─┐
q_1 -> 0 ┤ R(π/2,3π/2) ├─■─┤ R(π/2,5π/2) ├────────────────░────┤M├───
         ├─────────────┤ │ └─────────────┘                ░ ┌─┐└╥┘
q_0 -> 1 ┤ R(π/2,3π/2) ├─■────────■───────────────────────░─┤M├─╫────
         ├─────────────┤          │       ┌─────────────┐ ░ └╥┘ ║ ┌─┐
q_2 -> 2 ┤ R(π/2,3π/2) ├──────────■───────┤ R(π/2,5π/2) ├─░──╫──╫─┤M├
         └─────────────┘                  └─────────────┘ ░  ║  ║ └╥┘
 meas: 3/════════════════════════════════════════════════════╩══╩══╩═
                                                             0  1  2

Note that if you do this, you do need to provide the IQMBackend.run() method a qubit mapping that matches the restriction:

qubit_mapping = {i: backend.index_to_qubit_name(q) for i, q in enumerate(qubits)}
job = backend.run(transpiled_circuit, qubit_mapping=qubit_mapping)

Using custom IQM transpiler plugins#

For the native integration of the custom IQM transpiler passes with the Qiskit transpiler, we have implemented several scheduling plugins for the Qiskit transpiler. These plugins can be used as the scheduling_method string argument for transpile(). The mapping between these strings and the classes that implement the plugins is defined in the pyproject.toml file of this package. The documentation of these plugins in found in the respective plugin classes.

If you are unsure which plugin to use, you can use transpile_to_IQM() with the appropriate arguments. This function determines which plugin to use based on the backend and the provided arguments. Note that the Qiskit transpiler automatically uses the IQMDefaultSchedulingPlugin when the backend is an IQMBackend.

Noisy simulation of quantum circuit execution#

The execution of circuits can be simulated locally, with a noise model to mimic the real hardware as much as possible. To this end, Qiskit on IQM provides the class IQMFakeBackend that can be instantiated with properties of a certain QPU, e.g. using functions such as IQMFakeAdonis(), IQMFakeApollo() and IQMFakeAphrodite() that represent specific IQM quantum architectures with pre-defined, representative noise models.

from qiskit import transpile, QuantumCircuit
from iqm.qiskit_iqm import IQMFakeAdonis

circuit = QuantumCircuit(2)
circuit.h(0)
circuit.cx(0, 1)
circuit.measure_all()

backend = IQMFakeAdonis()
transpiled_circuit = transpile(circuit, backend=backend)
job = backend.run(transpiled_circuit, shots=1000)
print(job.result().get_counts())

Above, we use an IQMFakeAdonis() instance to run a noisy simulation of circuit on a simulated 5-qubit Adonis chip. The noise model includes relaxation ($T_1$) and dephasing ($T_2$), gate infidelities and readout errors. If you want to customize the noise model instead of using the default one provided by IQMFakeAdonis(), you can create a copy of the IQMFakeBackend instance with an updated error profile:

error_profile = backend.error_profile
error_profile.t1s['QB2'] = 30000.0  # Change T1 time of QB2 as example
custom_fake_backend = backend.copy_with_error_profile(error_profile)

Running a quantum circuit on a facade backend#

Circuits can be executed against a mock environment: an IQM server that has no real quantum computer hardware. Results from such executions are random bits. This may be useful when developing and testing software integrations.

Qiskit on IQM contains IQMFacadeBackend, which allows to combine the mock remote execution with a local noisy quantum circuit simulation. This way you can both validate your integration as well as get an idea of the expected circuit execution results.

To run a circuit this way, use the "facade_adonis" backend retrieved from the provider. Note that the provider must be initialized with the URL of a quantum computer with the equivalent architecture (i.e. names of qubits, their connectivity, and the native gateset should match the 5-qubit Adonis architecture).

from qiskit import transpile, QuantumCircuit
from iqm.qiskit_iqm import IQMProvider

circuit = QuantumCircuit(2)
circuit.h(0)
circuit.cx(0, 1)
circuit.measure_all()

iqm_server_url = "https://demo.qc.iqm.fi/cocos/"  # Replace this with the correct URL
provider = IQMProvider(iqm_server_url)
backend = provider.get_backend('facade_adonis')
transpiled_circuit = transpile(circuit, backend=backend)
job = backend.run(transpiled_circuit, shots=1000)
print(job.result().get_counts())