DeviceMetadata docs update. #4963
Conversation
MichaelBroughton
requested review from
cduck,
verult,
vtomole,
wcourtney and
a team
as code owners
| def qubit_set(self) -> FrozenSet['cirq.Qid']: | ||
| """Returns a set of qubits on the device, if possible. | ||
| """Returns the set of qubits on the device. |
There was a problem hiding this comment.
So objects like _UnconstrainedDevice would no longer be possible to write?
There was a problem hiding this comment.
We can still have the UnconstrainedDevice as it exists today. We just won't be able to add the metadata property to it because it doesn't have any qubits or nx_graph.
| if working_gatefamilies != gateset.gates: | ||
| missing_items = working_gatefamilies.difference(gateset.gates) | ||
| raise ValueError( | ||
| "Supplied gate_durations contains gates not present" | ||
| f" in supported_gates. {missing_items} in supported_gates" | ||
| f" in gateset. {missing_items} in gateset" | ||
| " is False." | ||
| ) |
There was a problem hiding this comment.
We are checking for equality but reporting difference in the ValueError. It's possible that working_gatefamilies is a subset of gateset.gates, in which case missing_items will be an empty set and error message would not be useful.
| "Supplied gate_durations contains gates not present" | ||
| " in gateset." |
There was a problem hiding this comment.
nit: This is still not correct, because if gate_durations is a subset of gateset, it won't contain any gates not present in the gateset, but vice versa would be true.
There was a problem hiding this comment.
Updated wording here to reflect the fact we are doing strict equality checks.
There was a problem hiding this comment.
gate_durations.keys should be allowed to be a subset of the gateset, because a gateset could contain GateFamilies that don't represent actual gates, such as FSimGateFamily.
An alternative Gateset design is to separate out GateFamilies like FSimGateFamily into a separate property equivalence_gate_families, which are used somewhat differently compared to GateFamilies representing gates.
MichaelBroughton
added
the
automerge
CirqBot
added
the
front_of_queue_automerge
CirqBot
removed
automerge
… of gateset (#5309) Fix for #4963 (comment) In the current Gateset design, some GateFamilies in a Gateset may not correspond to an actual device gate (e.g. `cirq_google.FSimGateFamily`). The new `GridDevice` implementation includes `FSimGateFamily` in its gateset to accept all equivalent Cirq representations of valid 2-qubit gates. @MichaelBroughton cc @tanujkhattar
… of gateset (quantumlib#5309) Fix for quantumlib#4963 (comment) In the current Gateset design, some GateFamilies in a Gateset may not correspond to an actual device gate (e.g. `cirq_google.FSimGateFamily`). The new `GridDevice` implementation includes `FSimGateFamily` in its gateset to accept all equivalent Cirq representations of valid 2-qubit gates. @MichaelBroughton cc @tanujkhattar
… of gateset (quantumlib#5309) Fix for quantumlib#4963 (comment) In the current Gateset design, some GateFamilies in a Gateset may not correspond to an actual device gate (e.g. `cirq_google.FSimGateFamily`). The new `GridDevice` implementation includes `FSimGateFamily` in its gateset to accept all equivalent Cirq representations of valid 2-qubit gates. @MichaelBroughton cc @tanujkhattar
Closes #4962