Implement num_qubits and num_clbits in Rust
#12495
Conversation
* QuantumCircuit.num_qubits and num_clbits are twice as fast after this PR. * These are used in several places. For example QuantumCircuit.width() is three times faster. * num_qubits and num_clbits are introduced in circuit_data.rs. These functions are called by the corresponding Python methods. They are also used in circuit_data.rs itself.
|
One or more of the following people are relevant to this code:
|
Pull Request Test Coverage Report for Build 9371133565Warning: This coverage report may be inaccurate.This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.
Details
💛 - Coveralls |
There was a problem hiding this comment.
This LGTM, in general as the source of truth moves more to the rust side I think this is fine.
There is an open question about a rust native representation of bits after #12459 but in the meantime asserting the bit pyobject vec in the circuit data is the source of truth seems fine to me (which this PR is doing implicitly), especially once we move to the gate representation being owned by the rust domain.
| @@ -220,6 +220,11 @@ impl CircuitData { | |||
| self.qubits.clone_ref(py) | |||
| } | |||
|
|
|||
| #[getter] | |||
There was a problem hiding this comment.
Do you want to throw a docstring on this and the other new methods just for completeness? While we don't publish docs for this as a pyclass (since it's internal only), it sometimes is a nice hint for people using IDEs that show the __doc__ attribute of a python class.
There was a problem hiding this comment.
Oh yeah. I forgot to add some of this when I moved it out of draft.
It wasn't 100% clear to me that qiskit/crates/circuit/src/circuit_data.rs Lines 151 to 152 in 797bb28 made it seem more likely.
Unfortunately, I expect there will be a lot of churn like this is gradually moving things to Rust. But this PR is pretty light in terms of changed code, so not too onerous to replace. |
My mistake tou're right it already is, I was misremembering the state of the code. The Python side |
| self.clbits_native.len() | ||
| } | ||
|
|
||
| pub fn width(&self) -> usize { |
There was a problem hiding this comment.
This is missing a docstring too.
mtreinish
added
performance
Changelog: None
* Implement num_qubits and num_clbits in Rust * QuantumCircuit.num_qubits and num_clbits are twice as fast after this PR. * These are used in several places. For example QuantumCircuit.width() is three times faster. * num_qubits and num_clbits are introduced in circuit_data.rs. These functions are called by the corresponding Python methods. They are also used in circuit_data.rs itself. * Add doc strings for rust implemented num_qubits and num_clbits * Add doc string for `width` in Rust
* Implement num_qubits and num_clbits in Rust * QuantumCircuit.num_qubits and num_clbits are twice as fast after this PR. * These are used in several places. For example QuantumCircuit.width() is three times faster. * num_qubits and num_clbits are introduced in circuit_data.rs. These functions are called by the corresponding Python methods. They are also used in circuit_data.rs itself. * Add doc strings for rust implemented num_qubits and num_clbits * Add doc string for `width` in Rust
QuantumCircuit.num_qubitsandnum_clbitsare twice as fast after this PR.QuantumCircuit.width()is three times faster.num_qubitsandnum_clbitsare introduced in circuit_data.rs. These functions are called by the corresponding Python methods. They are also used in circuit_data.rs itself.