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.

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

Utilities for approximate hardware noise #4671

Merged
merged 5 commits into from
Dec 13, 2021
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions cirq-core/cirq/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -88,6 +88,7 @@
NO_NOISE,
NOISE_MODEL_LIKE,
NoiseModel,
OpIdentifier,
SymmetricalQidPair,
UNCONSTRAINED_DEVICE,
NamedTopology,
Expand Down
12 changes: 12 additions & 0 deletions cirq-core/cirq/devices/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -47,3 +47,15 @@
get_placements,
draw_placements,
)

from cirq.devices.noise_utils import (
OpIdentifier,
decay_constant_to_xeb_fidelity,
decay_constant_to_pauli_error,
pauli_error_to_decay_constant,
xeb_fidelity_to_decay_constant,
pauli_error_from_t1,
pauli_error_from_depolarization,
average_error,
decoherence_pauli_error,
)
233 changes: 233 additions & 0 deletions cirq-core/cirq/devices/noise_utils.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,233 @@
# Copyright 2021 The Cirq Developers
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from typing import TYPE_CHECKING, Any, Dict, Tuple, Type, Union
import warnings
import numpy as np

from cirq import ops, protocols, value

if TYPE_CHECKING:
import cirq


# Tag for gates to which noise must be applied.
PHYSICAL_GATE_TAG = 'physical_gate'


@value.value_equality(distinct_child_types=True)
class OpIdentifier:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this class really needed ? (having looked at other PRs as well).

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We could probably get away without it, but the result would be a lot messier. Here are the alternatives I considered:

  • Just use Operation (it's hashable and contains the required info): this requires every current OpIdentifier reference to have some concept of "canonical" gates for each type supported in noise model generation (e.g. the PhasedXZGate constructor requires (x|z|axis_phase)_exponent args). This is considerably less convenient than just converting OpIDs to Operations during serialization.
  • Use a (Gate, qubits) tuple: this works about the same as OpIdentifier, but it's ultimately less readable to be saying op_id[0] and op_id[1:] instead of op_id.gate_type and op_id.qubits

If you have any other recommendations, I'm happy to consider them, but my sense from working through this is that OpIdentifier adds readability for relatively little cost. I opted not to surface OpIdentifier at the cirq package level to avoid any potential confusion.

"""Identifies an operation by gate and (optionally) target qubits."""

def __init__(self, gate_type: Type['cirq.Gate'], *qubits: 'cirq.Qid'):
self._gate_type = gate_type
self._gate_family = ops.GateFamily(gate_type)
self._qubits: Tuple['cirq.Qid', ...] = tuple(qubits)

@property
def gate_type(self) -> Type['cirq.Gate']:
# set to a type during initialization, never modified
return self._gate_type

@property
def qubits(self) -> Tuple['cirq.Qid', ...]:
return self._qubits

def _predicate(self, *args, **kwargs):
return self._gate_family._predicate(*args, **kwargs)

def swapped(self):
return OpIdentifier(self.gate_type, *self.qubits[::-1])

def is_proper_subtype_of(self, op_id: 'OpIdentifier'):
"""Returns true if this is contained within op_id, but not equal to it.
If this returns true, (x in self) implies (x in op_id), but the reverse
implication does not hold. op_id must be more general than self (either
by accepting any qubits or having a more general gate type) for this
to return true.
"""
more_specific_qubits = self.qubits and not op_id.qubits
more_specific_gate = self.gate_type != op_id.gate_type and issubclass(
self.gate_type, op_id.gate_type
)
if more_specific_qubits:
return more_specific_gate or self.gate_type == op_id.gate_type
elif more_specific_gate:
return more_specific_qubits or self.qubits == op_id.qubits
else:
return False

def __contains__(self, item: Union[ops.Gate, ops.Operation]) -> bool:
if isinstance(item, ops.Gate):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What would we do if we want an OpIdentifier for an Operation without a gate?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

None of the models from noise required that, so I opted not to support it. Is there a particular case you have in mind?

(CircuitOperations are a definite complication as far as matching is concerned, but I don't expect them to be used as the match target for an OpIdentifier.)

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In my use cases I had custom Operations that were not constructed using a Gate. I suppose the user could always generate a placeholder Gate object to identify any such Operation.

I can't give any native cirq Operation where this may be an issue since I am not so familiar with all the native cirq ops. One place you may have issues is with measurement operations. I.e. the output of cirq.measure has a gate output that is key dependent.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That case should be OK - this only checks against qubits and gate types, both of which are defined explictly in the cirq.measure call. On a more general note, Cirq has been moving towards an "all operations have gates" structure (#4683), which should help with any remaining problem cases.

return (not self._qubits) and self._predicate(item)
return (
(not self.qubits or (item.qubits == self._qubits))
and item.gate is not None
and self._predicate(item.gate)
)

def __str__(self):
return f'{self.gate_type}{self.qubits}'

def __repr__(self) -> str:
fullname = f'{self.gate_type.__module__}.{self.gate_type.__qualname__}'
qubits = ', '.join(map(repr, self.qubits))
return f'cirq.devices.noise_utils.OpIdentifier({fullname}, {qubits})'

def _value_equality_values_(self) -> Any:
return (self.gate_type, self.qubits)

def _json_dict_(self) -> Dict[str, Any]:
gate_json = protocols.json_cirq_type(self._gate_type)
return {
'gate_type': gate_json,
'qubits': self._qubits,
}

@classmethod
def _from_json_dict_(cls, gate_type, qubits, **kwargs) -> 'OpIdentifier':
gate_type = protocols.cirq_type_from_json(gate_type)
return cls(gate_type, *qubits)


# TODO: expose all from top-level cirq?
def decay_constant_to_xeb_fidelity(decay_constant: float, num_qubits: int = 2) -> float:
"""Calculates the XEB fidelity from the depolarization decay constant.
Args:
95-martin-orion marked this conversation as resolved.
Show resolved Hide resolved
decay_constant: Depolarization decay constant.
num_qubits: Number of qubits.
Returns:
Calculated XEB fidelity.
"""
N = 2 ** num_qubits
return 1 - ((1 - decay_constant) * (1 - 1 / N))


def decay_constant_to_pauli_error(decay_constant: float, num_qubits: int = 1) -> float:
"""Calculates pauli error from the depolarization decay constant.
Args:
decay_constant: Depolarization decay constant.
num_qubits: Number of qubits.
Returns:
Calculated Pauli error.
"""
N = 2 ** num_qubits
return (1 - decay_constant) * (1 - 1 / N / N)


def pauli_error_to_decay_constant(pauli_error: float, num_qubits: int = 1) -> float:
"""Calculates depolarization decay constant from pauli error.
Args:
pauli_error: The pauli error.
num_qubits: Number of qubits.
Returns:
Calculated depolarization decay constant.
"""
N = 2 ** num_qubits
return 1 - (pauli_error / (1 - 1 / N / N))


def xeb_fidelity_to_decay_constant(xeb_fidelity: float, num_qubits: int = 2) -> float:
"""Calculates the depolarization decay constant from XEB fidelity.
Args:
xeb_fidelity: The XEB fidelity.
num_qubits: Number of qubits.
Returns:
Calculated depolarization decay constant.
"""
N = 2 ** num_qubits
return 1 - (1 - xeb_fidelity) / (1 - 1 / N)


def pauli_error_from_t1(t_ns: float, t1_ns: float) -> float:
"""Calculates the pauli error from T1 decay constant.
This computes error for a specific duration, `t`.
Args:
t_ns: The duration of the gate in ns.
t1_ns: The T1 decay constant in ns.
Returns:
Calculated Pauli error resulting from T1 decay.
"""
t2 = 2 * t1_ns
return (1 - np.exp(-t_ns / t2)) / 2 + (1 - np.exp(-t_ns / t1_ns)) / 4


def pauli_error_from_depolarization(t_ns: float, t1_ns: float, pauli_error: float = 0) -> float:
"""Calculates the amount of pauli error from depolarization.
This computes non-T1 error for a specific duration, `t`. If pauli error
from T1 decay is more than total pauli error, this returns zero; otherwise,
it returns the portion of pauli error not attributable to T1 error.
Args:
t_ns: The duration of the gate in ns.
t1_ns: The T1 decay constant in ns.
pauli_error: The total pauli error.
Returns:
Calculated Pauli error resulting from depolarization.
"""
t1_pauli_error = pauli_error_from_t1(t_ns, t1_ns)
if pauli_error >= t1_pauli_error:
return pauli_error - t1_pauli_error

warnings.warn("Pauli error from T1 decay is greater than total Pauli error", RuntimeWarning)
return 0


def average_error(decay_constant: float, num_qubits: int = 1) -> float:
"""Calculates the average error from the depolarization decay constant.
Args:
decay_constant: Depolarization decay constant.
num_qubits: Number of qubits.
Returns:
Calculated average error.
"""
N = 2 ** num_qubits
return (1 - decay_constant) * (1 - 1 / N)


def decoherence_pauli_error(t1_ns: float, tphi_ns: float, gate_time_ns: float) -> float:
"""The component of Pauli error caused by decoherence.
Args:
t1_ns: T1 time in nanoseconds.
tphi_ns: Tphi time in nanoseconds.
gate_time_ns: Duration in nanoseconds of the gate affected by this error.
Returns:
Calculated Pauli error resulting from decoherence.
"""
gamma_2 = (1 / (2 * t1_ns)) + 1 / tphi_ns

exp1 = np.exp(-gate_time_ns / t1_ns)
exp2 = np.exp(-gate_time_ns * gamma_2)
px = 0.25 * (1 - exp1)
py = px
pz = 0.5 * (1 - exp2) - px
return px + py + pz
Loading