Qiskit / qiskit / 22931140556
88%
main: 88%

LAST BUILD BRANCH: gh-readonly-queue/main/pr-15635-31a2a1452dcf2b66ca42f567293217656d804a36
DEFAULT BRANCH: main
Ran
Jobs 1
Files 905
Run time 2min
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
coverage: 87.702% (-0.02%) from 87.719%
22931140556

push

github

web-flow 
Add hwb12 benchmarks to asv and pgo (#15792)

This commit adds the hwb12 benchmark from the reverseible logic
synthesis benchmarks [1]. The hwb12 benchmark is 12 qubit hidden
weighted bit function [2]. This is used in benchpress [3] as part of
the feynman toolkit's [4] benchmark suite and is the slowest benchmark
as part of benchpress's device benchmarks which run the transpiler on
this circuit targetting ibm_torino. Since Qiskit in particular is
relatively slow on this benchmark we should be tracking it as part of
our benchmkars. This adds it to the utility benchmarks although it's
technically not a "utility scale" circuit as it's only 20 qubits (12
plus ancilla/"garbage" qubits) because this has a nice harness in place
for benchmarking larger scale qasm circuits through the transpiler.
Additionally,since this is a particularly slow benchmark it adds it to
the set of circuits we run to collect profiling data on for PGO in the
hopes of speeding this up via compiler optimization.

One concern in particular is that Qiskit takes > 30 seconds  to locally
compile this circuit in the benchmark on the CZ backend. In general we'd
avoid adding such a slow benchmark. But because this is proving to be an
outlier in Qiskit's transpiler performance it's worth the runtime cost.
Especially given our general speed of execution in the asv benchmarks
currently this shouldn't be a blocker for getting things finished in the
time window on the benchmark system.

[1] https://reversiblebenchmarks.github.io/
[2] https://reversiblebenchmarks.github.io/hwbpoly.html
[3] https://github.com/Qiskit/benchpress/
[4] https://github.com/meamy/feynman

101148 of 115332 relevant lines covered (87.7%)

1165994.65 hits per line

Uncovered Existing Lines

Lines Coverage ∆ File
1
93.82
 -0.24% crates/qasm2/src/expr.rs
2
74.22
 0.2% crates/circuit/src/parameter/symbol_expr.rs
5
92.03
 -0.26% crates/qasm2/src/lex.rs
24
95.73
 -1.9% crates/qasm2/src/parse.rs
Jobs
ID Job ID Ran Files Coverage
1 22931140556.1 905
87.7
 GitHub Action Run
Source Files on build 22931140556