Qiskit
/
qiskit
/
15301926529
88%
main: 88%
main: 88%
|
Ran
|
Jobs
1
|
Files
789
|
Run time
2min
|
Badge
Embed ▾
README BADGES
|
coverage: 88.133% (+0.02%) from 88.11%
push
github
web-flowAvoid all-to-all connectivity in slowest test (#14483) (#14485) The current slowest unit test in the suite is: test_transpile_does_not_affect_backend_coupling_4_3 which is just sanity checking that a transpile() call doesn't mutate the CouplingMap by mistake. This is is a regression test for an old bug where that happened. However, in the move to use GenericBackendV2 it was creating a 130 qubit target with explicit all-to-all connectivity, which is the default behavior if you don't specify a coupling map argument. This ends up bogging the transpiler down in VF2PostLayout in optimization level 3 because there are a huge number of possible layouts, basically any permutation of 130 qubits. This isn't functionally part of the test because it's just trying to test if the coupling map is mutated by the transpiler. This commit updates the test to use a 130 qubit linear connectivity graph which still tests the circuit but takes a fraction of the time to execute because the target is much more constrained. (cherry picked from commit 88ce515da) Co-authored-by: Matthew Treinish <mtreinish@kortar.org>
72941 of 82762 relevant lines covered (88.13%)
347121.69 hits per line
Uncovered Existing Lines
| Lines | Coverage | ∆ | File |
|---|---|---|---|
| 1 |
94.81 |
0.0% | crates/accelerate/src/unitary_synthesis.rs |
| 1 |
94.23 |
-0.23% | crates/qasm2/src/expr.rs |
| 1 |
91.98 |
0.5% | crates/qasm2/src/lex.rs |
Jobs
| ID | Job ID | Ran | Files | Coverage | |
|---|---|---|---|---|---|
| 1 | 15301926529.1 | 789 |
88.13 |
GitHub Action Run |
