Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Cross Resonance Routine #821

Open
sorewachigauyo opened this issue May 3, 2024 · 11 comments · May be fixed by #860
Open

Cross Resonance Routine #821

sorewachigauyo opened this issue May 3, 2024 · 11 comments · May be fixed by #860
Assignees
@andrea-pasquale
Labels
urgent

Comments

@sorewachigauyo
Copy link

To implement a basic routine for the calibration of cross-resonance (CR) gates.

This involves sending a pulse on the control qubit with the frequency of the target qubit. We vary the pulse length to observe a Rabi oscillation. Based on the state of the control qubit, there will be a splitting of the Rabi frequencies, and we can select a pulse length $t$ such that the phase difference between the two oscillations is $\pi$. This then implements a CNOT gate.

image

A basic CR calibration is as follows:

sweep = np.arange(0, 1000, 20)
res1 = np.zeros(len(sweep))
res2 = np.zeros(len(sweep))

crtl_pi_pulse = platform.create_RX_pulse(qubit=CRTL_QUBIT, start=5)
tgt_ro_pulse = platform.create_qubit_readout_pulse(qubit=TGT_QUBIT, start=crtl_pi_pulse.finish)
cr_pulse = platform.create_RX_pulse(qubit=TGT_QUBIT, start=crtl_pi_pulse.finish)
cr_pulse.channel = crtl_pi_pulse.channel
cr_pulse.amplitude = 1

# We first perform a CR sweep when the control qubit is at the |0> state.
# This is equivalent to a typical Rabi length experiment
ps = PulseSequence(*[cr_pulse, tgt_ro_pulse])
for idx, t in enumerate(sweep):
    cr_pulse.duration = t
    tgt_ro_pulse.start = cr_pulse.finish
    res1[idx] = platform.execute_pulse_sequence(ps, opts)[tgt_ro_pulse.serial].magnitude

# Next, we add a pi-pulse on the control qubit at the start to get results for the |1> state.
ps.add(crtl_pi_pulse)
for idx, t in enumerate(sweep):
    cr_pulse.duration = t
    tgt_ro_pulse.start = cr_pulse.finish
    res2[idx] = platform.execute_pulse_sequence(ps, opts)[tgt_ro_pulse.serial].magnitude

The frequency of each Rabi oscillation is proportional to the amplitude of cr_pulse, so to obtain an optimal parameter for the amplitude, a 2D chevron of CR pulse amplitude and length can be performed.

sweeper_amplitude = Sweeper(
    Parameter.amplitude,
    amplitude_range,
    pulses=[cr_pulse],
    type=SweeperType.FACTOR,
)
sweeper_duration = Sweeper(
    Parameter.duration,
    duration_range,
    pulses=[cr_pulse],
    type=SweeperType.ABSOLUTE,
)
results = platform.sweep(
    ps,
    opts,
    sweeper_duration,
    sweeper_amplitude,
)
# and repeat for the other state of the control qubit

By also simultaneously driving the target qubit on resonance, an approximate interaction forms
$A_{CR} (\omega_{ZX} ZX + \omega_{IX} IX) + A_t \omega_{R2} IX $
where $A_{CR}$ is the amplitude of cr_pulse, $\omega_{ZX}, \omega_{IX}$ is determined by the target qubit Rabi frequency $\omega_{R2}$, effective qubit-qubit coupling, detunings and anharmonicites and $A_t$ is the amplitude of the second drive.

A basic experiment with the second drive is as follows:

crtl_pi_pulse = platform.create_RX_pulse(qubit=CRTL_QUBIT, start=5)
tgt_ro_pulse = platform.create_qubit_readout_pulse(qubit=TGT_QUBIT, start=crtl_pi_pulse.finish)
cr_pulse = platform.create_RX_pulse(qubit=TGT_QUBIT, start=crtl_pi_pulse.finish)
#cr_pulse.relative_phase = np.pi
IX_pulse = platform.create_RX_pulse(qubit=TGT_QUBIT, start=crtl_pi_pulse.finish)
IX_pulse.amplitude = 0.4
cr_pulse.channel = crtl_pi_pulse.channel
cr_pulse.amplitude = 1

ps = PulseSequence(*[cr_pulse, tgt_ro_pulse, IX_pulse])
sweeper_duration = Sweeper(
    Parameter.duration,
    duration_range,
    pulses=[cr_pulse, IX_pulse],
    type=SweeperType.ABSOLUTE,
)
results_crtl_gnd = platform.sweep(
    ps,
    opts,
    sweeper_duration
)
ps.add(crtl_pi_pulse)
results_crtl_exc = platform.sweep(
    ps,
    opts,
    sweeper_duration
)
@aorgazf
Copy link
Member

aorgazf commented May 15, 2024

Thanks @sorewachigauyo for raising this. It is also needed for the upcoming batch of TII's 2q chips.

@aorgazf aorgazf added the urgent label May 15, 2024
@andrea-pasquale andrea-pasquale linked a pull request May 19, 2024 that will close this issue
Draft
4 tasks
@jevillegasd
Copy link

I've added a version using sequences in 8597df3, @andrea-pasquale could you check if that code makes sense?

Even using the for-loop variation I cannot see much happening in these qubits
image
(the signal value is that of the ground state amplitude)

These are the qubits being used to run this test:

            "0": {
                "RX": {"duration": 80,"amplitude": 0.3745,"shape": "Gaussian(5)",
                    "frequency": 5312263656,"relative_start": 0,"phase": 0,"type": "qd"},
                "MZ": {
                    "duration": 2000,"amplitude": 0.15,"shape": "Rectangular()",
                    "frequency": 6910930000,"relative_start": 0,"phase": 0,"type": "ro"
                }
            },
            "1": {
                "RX": {"duration": 80,"amplitude":  0.5448,"shape": "Gaussian(5)",
                    "frequency": 5393450614,"relative_start": 0,"phase": 0,"type": "qd"},
                "MZ": {
                    "duration": 2000,"amplitude": 0.1,"shape": "Rectangular()",
                    "frequency": 7.02001e+09,"relative_start": 0,"phase": 0,"type": "ro"
                }
            },

@sorewachigauyo
Copy link
Author

sorewachigauyo commented Sep 17, 2024 •
edited
Loading

Hi @jevillegasd,

If its not too sensitive, could you give some more information about the qubit system you are using? Given that the qubits are 80MHz apart, it is strange that you do not see an effect or even crosstalk-driven Rabi oscillations.

We can take it to emails if needed.

Maybe you need more power either with an amplifier or higher amplitude.

@jevillegasd
Copy link

Absolutly the QPU is a three qubit chip with fixed frequency qubits coupled together in a straight line (you can see the platform here qiboteam/qibolab_platforms_qrc@77b7b32). The chip has an additional witness qubit and a resonator. Here I list the most relevant parameters.

  f_ro hp , GHz f_ro lp , GHz $\chi$ , MHz f_01 , GHz f_02/2 , GHz T1 [us] T2 [us]
Q1 6.9111 6.9060 5.03 5.312 5.150 6.3 7.1
Q2 7.020138 7.015958 4.17 5.393 5.242 4.5 3.6
Q3 7.140432 7.135537 4.89 5.913 5.764 5.2 5.4
QW 7.626598 7.621588 5.01 4.908 4.768    
Free Res 7.892598 7.892598 0        

@jevillegasd
Copy link

I've run the routine using [0,0] as target and control qubits, which should just make a normal Rabi, this to confirm if its working. And here is what I get

image
which is arguably good.

I will try increasing the power and see where that takes us.

@Edoardo-Pedicillo
Copy link
Contributor

I will try increasing the power and see where that takes us.

Have you tried to run a cross resonance chevron ?

@sorewachigauyo
Copy link
Author

The qubits are coupled directly or by bus resonator? Do you have an estimate of the coupling strength from designed values or measured by a two-qubit simultaneous on-resonance Ramsey?

@jevillegasd
Copy link

I will try increasing the power and see where that takes us.

Have you tried to run a cross resonance chevron ?

I still need to make the version with sequences for that as well so that I can use it with QBlox.

@jevillegasd
Copy link

The qubits are coupled directly or by bus resonator? Do you have an estimate of the coupling strength from designed values or measured by a two-qubit simultaneous on-resonance Ramsey?

They are directly coupled with a transmission line with a coupling capacitance of 0.5 fF at each transmon, this is about 10 MHz by design. We have not tried to extract this value using simultaneous Ramsey tests, could you explain how can we get this?

@andrea-pasquale
Copy link
Contributor

The qubits are coupled directly or by bus resonator? Do you have an estimate of the coupling strength from designed values or measured by a two-qubit simultaneous on-resonance Ramsey?

They are directly coupled with a transmission line with a coupling capacitance of 0.5 fF at each transmon, this is about 10 MHz by design. We have not tried to extract this value using simultaneous Ramsey tests, could you explain how can we get this?

If @sorewachigauyo is referring to the ZZ coupling, there is a PR where we have a draft for that protocol #970.
Essentially if you do a Ramsey experiment on one qubit will keeping another qubit in state 1, if you compare it with the original ramsey from the phase shift in the oscillations you should be able to get the ZZ coupling.
However, we are using again duration sweepers so it will not work out of the box with qblox.

@sorewachigauyo
Copy link
Author

sorewachigauyo commented Sep 18, 2024 •
edited
Loading

Yeah, #970 should help you determine the coupling strength and check if the qubits are coupled or not. Based on the coherence times you measured, you should observe the Ramsey oscillations within the coherence.

@jevillegasdatTII jevillegasdatTII mentioned this issue Sep 27, 2024
Closed
8 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@andrea-pasquale andrea-pasquale

Labels
urgent
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Cross resonance protocols qiboteam/qibocal
5 participants
@jevillegasd @sorewachigauyo @andrea-pasquale @aorgazf @Edoardo-Pedicillo