GitLab CI
c3589552d2
perf: Update aggregated performance analysis [ci skip]
2026-01-23 16:16:33 -05:00
6d4b0aab2e
fix: stop pool metrics collector when tests finish to avoid skewed averages
2026-01-23 15:05:07 -05:00
08dbc8ff75
chore: bump version to 4.2.18
2026-01-23 10:42:50 -05:00
5c14ccc3af
perf: regenerate aggregate with just 4.2.11 and 4.2.12 baseline
2026-01-23 08:52:34 -05:00
GitLab CI
4087339356
perf: Update aggregated performance analysis [ci skip]
2026-01-23 08:30:52 -05:00
c5510100da
chore: remove old aggregated report for fresh baseline
2026-01-23 08:20:38 -05:00
GitLab CI
d87401905b
perf: Update aggregated performance analysis [ci skip]
2026-01-23 07:14:31 -05:00
GitLab CI
06cd70d1ed
perf: Update aggregated performance analysis [ci skip]
2026-01-23 06:46:38 -05:00
d17f326f34
fix: resolve merge conflict markers in aggregated performance report
2026-01-23 05:51:41 -05:00
GitLab CI
d3e2c99111
perf: Update aggregated performance analysis [ci skip]
2026-01-23 05:05:55 -05:00
GitLab CI
1e6f06845c
perf: Update aggregated performance analysis [ci skip]
2026-01-23 05:01:59 -05:00
GitLab CI
646d980f08
perf: Update aggregated performance analysis [ci skip]
2026-01-23 04:36:41 -05:00
b37ce346d2
Update aggregated performance analysis with 4.2.6 release
...
Now analyzing 5 releases: 4.2.0, 4.2.3, 4.2.4, 4.2.5, 4.2.6
Key findings with 4.2.6 included:
- Haskell now shows 509.5% variance (21s → 128s)
- V shows 418.2% variance (22s → 114s)
- Elixir remains highly variable at 425.0% (20s → 105s)
- Average duration for 4.2.6: 54s (improvement from 4.2.5's 67s)
- Still shows non-deterministic scheduling patterns
Updated charts reflect 5-release trend analysis.
2026-01-19 15:24:52 -05:00
GitLab CI
12ea3f77f6
perf: Update aggregated performance analysis [ci skip]
2026-01-19 15:22:43 -05:00
1f83eaf175
Add aggregated performance analysis with dynamic version discovery
...
- Add perf-aggregate-report CI job to analyze variance across releases
- Implement dynamic version discovery using git tags (ever-growing)
- Generate charts via UN sandbox using matplotlib
- Create AGGREGATED-PERFORMANCE.md with comprehensive methodology
- Add Makefile target for local report generation
- Include 3 visualization charts showing variance trends
Key findings: 2-3x performance variance due to orchestrator placement
on CPU-bound pool causing non-deterministic scheduling.
2026-01-19 13:13:22 -05:00
