Qiskit / qiskit
88%
main: 88%

LAST BUILD BRANCH: circuit-annotations-spelling
DEFAULT BRANCH: main
Repo Added
Files 871
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

LAST BUILD ON BRANCH gh-readonly-queue/main/pr-15131-0357a770d2775f8d2998edd2857f02bc1997a432
branch: gh-readonly-queue/main/pr-15131-0357a770d2775f8d2998edd2857f02bc1997a432
CHANGE BRANCH
x

coverage: 88.237% (+0.02%) from 88.217%
18349671697

push

github

web-flow 
Check target gates are parameterized in optimize1qgatesdecomposition (#15131)

* Check target gates are parameterized in optimize1qgatesdecomposition

This commit updates the Optimize1qGatesDecomposition transpiler pass to
check the defined operation in the target to ensure it's a valid gate to
use for the decomposition. Previously the pass would just use the gate
name for each qubit and treat that as a basis gate list. However this
ignored the extra details that the target provides as a target gate can
match the name but not be valid for use in synthesis. This commit fixes
that oversight by checking the gates directly and ensuring they're valid
for use in the decomposition.

Fixes #14743

* Fix release note typos

Co-authored-by: Jake Lishman <jake@binhbar.com>

* Simplify logic in pass

Since this PR is intended to be backported to 2.2.x and 1.4.x we should
minimize the diff. The previous version of this PR changed the behavior
of the pass to normalize on using standard gates even if there were
custom names for gates in the target. Longer term I think this behavior
is better for the pass, but there is discussions around exactly the
semantics of the Target's data model around names vs operations here and
also that this might change the output pretty noticeably in some cases.
For a bugfix intended for backport it's better to be conservative here,
so this commit reverts the logic to just use the name in the target and
only filter operations if the target doesn't actually support arbitrary
angles on them. We can revisit the improvements in a follow up PR.

---------

Co-authored-by: Jake Lishman <jake@binhbar.com>

15 of 15 new or added lines in 1 file covered. (100.0%)

11 existing lines in 5 files now uncovered.

93142 of 105559 relevant lines covered (88.24%)

1160733.51 hits per line

Relevant lines Covered
105559 RELEVANT LINES 93142 COVERED LINES
1160733.51 HITS PER LINE
Source Files on gh-readonly-queue/main/pr-15131-0357a770d2775f8d2998edd2857f02bc1997a432

Recent builds

Builds Branch Commit Type Ran Committer Via Coverage
18349671697 gh-readonly-queue/main/pr-15131-0357a770d2775f8d2998edd2857f02bc1997a432 Check target gates are parameterized in optimize1qgatesdecomposition (#15131) * Check target gates are parameterized in optimize1qgatesdecomposition This commit updates the Optimize1qGatesDecomposition transpiler pass to check the defined operat... push web-flow github
88.24
See All Builds (14065)