hyperledger / fabric-x-committer / 29756083315 / 3
91%
main: 91%

LAST BUILD BRANCH: bench-report-metrics
DEFAULT BRANCH: main
Ran
Files 127
Run time 5s
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: 79.679% (-0.01%) from 79.691%
29756083315.3

Pull #688

github

liran-funaro 
[loadgen] Remove the workload distribution

- Improvement

- The load generator's `Distribution` descriptor supported five variants
(`const`, `uniform`, `normal`, `bernoulli`, `discrete`), but only `const`
was ever used. Load generation holds each parameter at a fixed value so
results are deterministic and comparable across runs, so the unused
variants only added config surface, sample-YAML documentation, generator
types, and tests with no realized benefit. This collapses `Distribution`
down to plain constant values.

- Delete `distributions.go` and the orphaned
`Const`/`FloatToInt`/`FloatToPositiveInt`/`FloatToBoolean` generators;
`MultiGenerator.Count` is now a plain `int`.
- Replace the `*Distribution` profile fields with plain scalars:
read/read-write/blind-write counts and query size become `uint32`, the
minimal invalid-keys portion becomes `Probability`, and the dependency
gap becomes `uint64` (min 1).
- Keep per-transaction conflict injection (invalid signatures and
dependency probability) via a small `bernoulliGenerator`
(`rnd.Float64() < p`) with identical semantics to the previous
Bernoulli-based path.
- Update the sample `loadgen.yaml` and the `loadgen_shared.yaml.tmpl`
config template to the plain-number syntax, plus the affected tests.

- Resolves #682

Signed-off-by: Liran Funaro <liran.funaro@gmail.com>
Pull Request #688: [loadgen] Remove the workload distribution

8983 of 11274 relevant lines covered (79.68%)

34353.86 hits per line

Source Files on job db-test - 29756083315.3