Move QuantumCircuit.assign_parameters to Rust (backport #12794)
#12878
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![mergify[bot]](https://avatars.githubusercontent.com/in/10562?s=60&v=4){kind=small}
* Move `QuantumCircuit.assign_parameters` to Rust This is (as far as I could tell), the last really major performance regression in our asv suite compared to 1.1.0, so with this commit, we should be at _not worse_ for important utility-scale benchmarks. This largely rewrites `ParamTable` (renamed back to `ParameterTable` because I kept getting confused with `Param`) to have more Rust-friendly interfaces available, so that `assign_parameters` can then use them. This represents a 2-3x speedup in `assign_parameters` performance over 1.1.0, when binding simple `Parameter` instances. Approximately 75% of the time is now spent in Python-space `Parameter.assign` and `ParameterExpression.numeric` calls; almost all of this could be removed were we to move `Parameter` and `ParameterExpression` to have their data exposed directly to Rust space. The percentage of time spent in Python space only increases if the expressions to be bound are actual `ParameterExpression`s and not just `Parameter`. Most changes in the test suite are because of the use of internal-only methods that changed with the new `ParameterTable`. The only discrepancy is a bug in `test_pauli_feature_map`, which was trying to assign using a set. * Add unit test of parameter insertion This catches a bug that was present in the parent commit, but this PR fixes. * Update crates/circuit/src/imports.rs * Fix assignment to `AnnotatedOperation` * Rename `CircuitData::num_params` to match normal terminology * Fix typos and 🇺🇸 * Fix lint (cherry picked from commit a68de4f)
|
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 following people are relevant to this code:
|
github-actions
bot
added
priority: high
performance
Changelog: None
github-actions
bot
added
the
mod: circuit
jakelishman
approved these changes
Aug 1, 2024
Pull Request Test Coverage Report for Build 10197698473Details
💛 - Coveralls |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This is (as far as I could tell), the last really major performance regression in our asv suite compared to 1.1.0, so with this commit, we should be at not worse for important utility-scale benchmarks.
This largely rewrites
ParamTable(renamed back toParameterTablebecause I kept getting confused withParam) to have more Rust-friendly interfaces available, so thatassign_parameterscan then use them.This represents a 2-3x speedup in
assign_parametersperformance over 1.1.0, when binding simpleParameterinstances. Approximately 75% of the time is now spent in Python-spaceParameter.assignandParameterExpression.numericcalls; almost all of this could be removed were we to moveParameterandParameterExpressionto have their data exposed directly to Rust space. The percentage of time spent in Python space only increases if the expressions to be bound are actualParameterExpressions and not justParameter.Most changes in the test suite are because of the use of internal-only methods that changed with the new
ParameterTable. The only discrepancy is a bug intest_pauli_feature_map, which was trying to assign using a set.Details and comments
Built on #12730, so will need rebasing over it.
I think this first commit might accidentally have introduced a small (10%) regression to parametric-circuit construction time over its parent. That's a mistake if so - I should be able to fix that later.edit: on retiming, I couldn't reproduce a problem - if anything, this commit is a minor improvement.Timings for parametric circuit benchmarks compared to 1.1.0 (the different SHA1 is because I hadn't written the commit message when I took the benchmark):