Converted T1 #277

omrieoqm · 2025-01-26T07:06:07Z

No description provided.

…ure' into better_note_structure_T1

deanpoulos

Hi Omrie, the refactoring looks great! I have been extra picky in my additional comments because I think it will have a benefit to how we write code and refactor nodes in the long run. There is also an important functional change that needs to be made to the behaviour of multiplexing.

deanpoulos · 2025-01-29T11:11:47Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

-    multiplexed: bool = False
-
-node = QualibrationNode(name="05_T1", parameters=Parameters())
+node = QualibrationNode(name="05_T1", parameters=Parameters(qubits = None,


Please indent it as follows:

QualibrationNode( name="05_T1", parameters=Parameters( qubits = None, num_averages = 100, min_wait_time_in_ns = 16, max_wait_time_in_ns = 100000, wait_time_step_in_ns = 600, flux_point_joint_or_independent_or_arbitrary = "independent", reset_type = "thermal", use_state_discrimination = False, simulate = False, simulation_duration_ns = 2500, timeout = 100, load_data_id = None, multiplexed = False ) )

This kind of nesting makes it clearer that name and parameters are attributes of the QualibrationNode, and that the parameters belong to node.parameters.

deanpoulos · 2025-01-29T11:13:24Z

Quantum-Control-Applications-QuAM/Superconducting/quam_libs/experiments/T1/parameters.py

+    flux_point_joint_or_independent_or_arbitrary: Literal["joint", "independent", "arbitrary"] = "independent"
+    reset_type: Literal["active", "thermal"] = "thermal"
+    use_state_discrimination: bool = False
+    simulate: bool = False


Here you have re-defined many of the parameters which will be inherited from the parent classes.

Please remove those which are duplicated, i.e.,

simulate: bool = False simulation_duration_ns: int = 2500 timeout: int = 100 load_data_id: Optional[int] = None multiplexed: bool = False use_waveform_report: bool = False flux_point_joint_or_independent_or_arbitrary: Literal["joint", "independent", "arbitrary"] = "independent"

And any others that I missed

You are right. I now see what you did there. It is nice.

deanpoulos · 2025-01-29T11:14:19Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

 config = machine.generate_config()
-# Open Communication with the QOP
 if node.parameters.load_data_id is None:
    qmm = machine.connect()


Please add a new line before the start of the # %% {QUA_program} section

deanpoulos · 2025-01-29T11:24:00Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

-    qubits = [machine.qubits[q] for q in node.parameters.qubits]
-num_qubits = len(qubits)
-
-
 # %% {QUA_program}
 n_avg = node.parameters.num_averages  # The number of averages
 # Dephasing time sweep (in clock cycles = 4ns) - minimum is 4 clock cycles


Here we create variables which will be used in the program:

idle_times

detunings

arb_flux_bias_offset

Glancing at how these variables are created doesn't really give me any useful information about what what they are doing and how they will be used. So, I think it should be put into a separate module.

The module I think it fits best in would be quam_libs/experiments/T1/parameters.py, specifically as a method to the Parameters class. But, due to this PR in QUAlibrate, methods of Parameters classes get removed. Instead, please make it into functions that live in the same module.

The outcome in the main node is that these values should be acquired in the following way:

from quam_libs.experiments.T1.parameters import Parameters ... idle_times = get_idle_times_in_clock_cycles(node.parameters) detunings = ... # this doesn't actually get used. Ask Tom if it's a bug that needs fixing, or should be removed # maybe think of a better name for this? It feels like 2 or 3 of the "flux", "bias", "offset" and "arb" are redundant arb_flux_bias_offset = get_arb_flux_bias_offset_for_each_qubit(node.parameters, qubits)

For reference, please see the refactored Ramsey node.

Done. Deleted detuning (checked with Tom). arb_flux_bias_offset changed to arb_flux_offset

deanpoulos · 2025-01-29T11:31:57Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

A note on the behaviour of the program- the behaviour when setting multiplexed = True is actually wrong. What we want is behaviour that is multiplexed and aligned.

Given how it's currently written, imagine that multiplexed=True, then imagine what would happen in any of the following scenarios:

active reset takes longer to "succeed" for q1 than q2 (if using active reset)

thermal reset takes longer for q1 than q2 due to different T1s (if using thermal reset)

the pi-pulse times are different between q1 and q2

the readout between q1 and q2 have different durations

If it isn't clear straight away, you should visualize this using the waveform report.

Please rewrite the program so that the readout is aligned for all qubits, in a way that ensures that the idle time is the same for all qubits. This might involve some sort of "right-aligning" for the pi-pulses if they have different durations which I'm not sure is possible. If it's not, just assume the pi-pulses have the same duration.

If it still isn't clear, reference the refactored Ramsey node or this section of the confluence page.

I made the proper changes. Let's discuss this.

deanpoulos · 2025-01-29T11:34:06Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

@@ -187,56 +160,13 @@ class Parameters(NodeParameters):

    # %% {Data_analysis}
    # Fit the exponential decay


Since you call a function called fit_exponential_decay, you definitely don't need a comment which says
# Fit the exponential decay. Please remove comments like this throughout the file.

It looks nicer to have "white space" like new lines for readability than having redundant comments

deanpoulos · 2025-01-29T11:37:42Z

Quantum-Control-Applications-QuAM/Superconducting/calibration_graph/05_T1.py

@@ -245,7 +175,6 @@ class Parameters(NodeParameters):
                and tau_error.sel(qubit=q.name).values / float(tau.sel(qubit=q.name).values) < 1


This logic isn't so clear to me, and it's hard to read at first glance. I'm sure it's the same for you and any other users. Could you spend a minute or two trying to make sense of it, then split the if-statement into something like this:

t1_is_positive = float(tau.sel(qubit=q.name).values) > 0 t1_XYZ = (tau_error.sel(qubit=q.name).values / float(tau.sel(qubit=q.name).values)) < 1 if (t1_is_positive and t1_is_XYZ): q.T1 = ...

Done. I think the goal is to check the error in the fit is smaller than the actual fit value. Probably done to verify that the fit make sense. I will ask Akiva.

deanpoulos · 2025-01-29T11:40:44Z

Quantum-Control-Applications-QuAM/Superconducting/quam_libs/components/quam_root.py

-            settings["port"] = self.network["port"]
-        self.qmm = QuantumMachinesManager(**settings)
+
+        if self.network["cloud"]:


This will break everyone's code unless they have a field called cloud in self.network.

In order to avoid this, the safer option would be:

if self.network.get("cloud", False)

This treats "cloud" as an optional parameter of the dictionary, which resolves to False if it's not present.

deanpoulos · 2025-01-29T11:41:58Z

Quantum-Control-Applications-QuAM/Superconducting/quam_libs/experiments/T1/analysis.py

+from quam_libs.lib.save_utils import fetch_results_as_xarray
+from quam_libs.lib.fit import fit_decay_exp, decay_exp
+
+def fetch_dataset(job: QmJob, qubits: List[Transmon], idle_times: np.ndarray, unit: unit) -> xr.Dataset:


I think it's clearer to exlclude the unit object as an argument to this function.

Either re-instantiate it in the function, or do the unit conversion manually since it is trivial.

omrieoqm added 5 commits January 23, 2025 16:35

converted T1

2f379b9

Merge remote-tracking branch 'origin/sketch/SCquam/better_node_struct…

1a776b6

…ure' into better_note_structure_T1

T1 convert

af195a5

clound connection

3164634

changed into to the get_qubit_used_in_node

d94fc65

omrieoqm requested a review from deanpoulos January 26, 2025 07:06

fixed indentation in connect

1a6d394

deanpoulos requested changes Jan 29, 2025

View reviewed changes

deanpoulos assigned deanpoulos and omrieoqm and unassigned deanpoulos Jan 29, 2025

omrieoqm added 2 commits January 30, 2025 10:21

Changes from Dean comments

8147aac

node testing. flux_point_joint_or_independent_or_arbitrary.

3313e5f

omrieoqm merged commit 7b7ea38 into sketch/SCquam/better_node_structure Jan 30, 2025
1 check failed

omrieoqm deleted the better_note_structure_T1 branch January 30, 2025 10:51

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Converted T1 #277

Converted T1 #277

omrieoqm commented Jan 26, 2025

deanpoulos left a comment

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

deanpoulos Jan 29, 2025

omrieoqm Jan 30, 2025

		@@ -187,56 +160,13 @@ class Parameters(NodeParameters):

		# %% {Data_analysis}
		# Fit the exponential decay

		@@ -245,7 +175,6 @@ class Parameters(NodeParameters):
		and tau_error.sel(qubit=q.name).values / float(tau.sel(qubit=q.name).values) < 1

Converted T1 #277

Converted T1 #277

Conversation

omrieoqm commented Jan 26, 2025

deanpoulos left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment