Readd support for backends with directed basis gates to RB experiments #960
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
nkanazawa1989
approved these changes
Nov 22, 2022
There was a problem hiding this comment.
Thanks Itoko-san. I think this is reasonable handling of directed gates with the minimum coding overhead for now. Unfortunately a typical IBM backend provides fake bidirectional gates with Hadamard trick, and this could increase the variance of RB samples because of infidelity of extra Hadamard gates. Eventually we should only consider the native direction gates (rather than fake bidirectional). When only one-way native direction is available, I think we can flip the gate direction but merge Hadamard gates into successor and predecessor single qubit gates.
I think this can be good to go now but we must come back to this at some point.
itoko
added a commit
to itoko/qiskit-experiments
that referenced
this pull request
Nov 29, 2022
qiskit-community#960) * Add tests for directed basis gates * Readd support of backends with only directed 2q-basis
itoko
added a commit
to itoko/qiskit-experiments
that referenced
this pull request
Nov 29, 2022
qiskit-community#960) * Add tests for directed basis gates * Readd support of backends with only directed 2q-basis
itoko
added a commit
that referenced
this pull request
Dec 23, 2022
Improve the performance of transpiled circuit generation in 1Q/2Q StandardRB/InterleavedRB (about 10x speedup in 1Q SRB/IRB and 5x speedup in 2Q SRB/IRB in most cases). Including following feature pull-requests: * New algorithm for RB building Cliffords by layers (#892) * Improve custom transpilation for faster 1Q/2Q RB (#922) * Improve integer-based Clifford operations for 1Q/2Q RB (#940) * Readd support for backends with directed basis gates to RB experiments (#960) Features other than speedup: * Fix performance regression in 3 or more qubits RB (embedded at Refactor RB module for future extensions #898) * Improve validation of interleave_element in InterleavedRB * Restructure tests for RB module * Remove the barriers in front of the first Cliffords in RB circuits Co-authored-by: merav-aharoni <46567124+merav-aharoni@users.noreply.github.com> Co-authored-by: Naoki Kanazawa <nkanazawa1989@gmail.com>
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 PR fixes a feature regression introduced in #922 by re-adding the support for backends with directed 2-qubit gates to RB experiments.
Details and comments
The feature was accidentally dropped at #922 that replaced the call of
transpile(circ, backend)withtranspile(circ, basis_gates=basis_gates). For example, 2Q StandardRB does not work for backends that it support cx(0, 1) but not support cx(1, 0).To reimplement the feature, this commit adds a fallback call of
transpile(circ, backend)in the case when the backend does not support any undirected 2-qubit gates on the physical qubits to be benchmarked.