Move utility functions _inverse_pattern and _get_ordered_swap to Rust

[jpacold]

Summary

This PR starts addressing #12225. I moved the implementations of two functions from qiskit/synthesis/permutation/permutation_utils.py to Rust.

Details and comments

• Created permutation.rs

  • I added some checks to ensure that inputs are actually permutations. Otherwise, _get_ordered_swap uses the same logic as it did in Python.
  • Performance is better than Python, but sensitive to input type (more notes below)

• Added a module qiskit._accelerate.permutation and a submodule qiskit._accelerate.permutation.permutation_utils (following the structure of the existing Python modules). The wrapped functions are in the permutation_utils submodule.

• Replaced the function bodies in permutation_utils.py with calls to the Rust functions.

• Added tests.

Performance

As expected, this implementation of _get_ordered_swap is faster than the Python version. However, the input type is expected to be an array of int64 and there is a performance cost for converting from any other type. On my machine:

Input type	Speedup over Python
Numpy array (int64)	~15x
Numpy array (int32, uint32, or uint64)	~4x
Python list	~7x

For now, I don't think this matters much, because the bottleneck in synth_permutation_basic is constructing the circuit, but it is something that might need to be optimized later on.

[CLAassistant]

All committers have signed the CLA.

[coveralls]

Pull Request Test Coverage Report for Build 9311713986

Details

• 78 of 78 (100.0%) changed or added relevant lines in 4 files are covered.
• 3 unchanged lines in 1 file lost coverage.
• Overall coverage increased (+0.03%) to 89.625%

---

Files with Coverage Reduction	New Missed Lines	%
crates/qasm2/src/lex.rs	3	93.13%

Totals
Change from base Build 9291000954:	0.03%
Covered Lines:	62399
Relevant Lines:	69622

---

💛 - Coveralls

[qiskit-bot]

Thank you for opening a new pull request.

Before your PR can be merged it will first need to pass continuous integration tests and be reviewed. Sometimes the review process can be slow, so please be patient.

While you're waiting, please feel free to review other open PRs. While only a subset of people are authorized to approve pull requests for merging, everyone is encouraged to review open pull requests. Doing reviews helps reduce the burden on the core team and helps make the project's code better for everyone.

One or more of the the following people are requested to review this:

• @Eric-Arellano
• @Qiskit/terra-core
• @kevinhartman
• @mtreinish

[raynelfss]

Is this still a work in progress? @jpacold if it's not, it would be helpful to remove the [WIP] from the title as it seems to me this has been completed.

[jpacold]

@raynelfss Done, thanks for pointing that out.

[raynelfss]

LGTM, just a small correction on the docstrings for your rust functions. Thank you for working on this!

[ShellyGarion]

why do you need two [pymodule]s here? I see that in other cases only one was used.

[jpacold]

It's not really necessary. I copied the structure from the Python synthesis module:

qiskit
    |-- synthesis
        |-- permutation
            | -- permutation_full
            | -- permutation_lnn
            | -- permutation_reverse_lnn
            | -- permutation_utils

since I was thinking the idea is to eventually move all of it to Rust.

[ShellyGarion]

you are right that the idea is to eventually move all of the synthesis code to rust,
however, it seems that currently the structure is flat, see e.g.

qiskit/qiskit/__init__.py

Line 83 in 6e3bc2c

sys.modules["qiskit._accelerate.two_qubit_decompose"] = qiskit._accelerate.two_qubit_decompose

qiskit/qiskit/__init__.py

Line 70 in 6e3bc2c

sys.modules["qiskit._accelerate.euler_one_qubit_decomposer"] = (

[jpacold]

Got it, I made another commit to flatten this as well.