Skip to content

docs: update incremental benchmarks (3.3.0)#526

Merged
carlos-alm merged 2 commits intomainfrom
benchmark/incremental-v3.3.0-20260319-081441
Mar 19, 2026
Merged

docs: update incremental benchmarks (3.3.0)#526
carlos-alm merged 2 commits intomainfrom
benchmark/incremental-v3.3.0-20260319-081441

Conversation

@github-actions
Copy link
Copy Markdown
Contributor

Automated incremental benchmark update for 3.3.0 from workflow run #381.

@greptile-apps
Copy link
Copy Markdown
Contributor

greptile-apps bot commented Mar 19, 2026

Greptile Summary

This is an automated documentation PR that records incremental benchmark results for v3.3.0, adding two new rows to the historical summary table (native + wasm engines), updating the "Latest results" section, and prepending a new JSON entry to the embedded machine-readable data block. All display values, raw JSON figures, and percentage-change deltas are internally consistent and mathematically correct.

Key observations from the new benchmark data:

  • Significant build-time regression: Native full build climbed from 1.3 s (v3.1.4, 398 files) to 3.0 s (v3.3.0, 429 files) — a 141 % increase despite only ~8 % more files. WASM full build also grew from 3.3 s → 4.3 s (+29 %).
  • Import-pair explosion: Tracked import pairs jumped from 175 → 664 (≈3.8×), which explains the corresponding 700–888 % increase in native/JS resolution times and appears to be the primary driver of the build-time regression.
  • Missing intermediate versions: The benchmark history jumps from v3.1.4 directly to v3.3.0, with no entries for any 3.2.x releases. This is worth noting if benchmarks are expected to be recorded for every release.
  • The per-import times (0 ms each) and speedup ratio (2.7×) are consistent with the raw figures.

Confidence Score: 5/5

  • This is a documentation-only, automated benchmark update with no code changes — safe to merge.
  • The PR contains a single Markdown file updated by a GitHub Actions bot. The data is internally consistent (JSON values match table display values and percentage deltas are mathematically correct). There are no logic, syntax, or security concerns. The only observations are informational: a notable build-time regression in the underlying codebase (visible in the numbers) and missing 3.2.x benchmark entries, neither of which are issues introduced by this PR.
  • No files require special attention.

Important Files Changed

Filename Overview
generated/benchmarks/INCREMENTAL-BENCHMARKS.md Automated benchmark data update for v3.3.0 — adds two new table rows (native + wasm), refreshes the "Latest results" section, and prepends a new JSON entry to the embedded data block. All display values match the raw JSON figures and percentage deltas are mathematically correct.

Sequence Diagram

sequenceDiagram
    participant W as GitHub Actions Workflow #381
    participant R as codegraph (self-measured)
    participant G as generated/benchmarks/INCREMENTAL-BENCHMARKS.md
    participant PR as Pull Request #526

    W->>R: Run incremental benchmark suite (v3.3.0)
    R->>R: Full build (native 3.0s, wasm 4.3s)
    R->>R: No-op rebuild (native 10ms, wasm 11ms)
    R->>R: 1-file rebuild (native 417ms, wasm 479ms)
    R->>R: Import resolution (664 pairs, 9.6ms native / 25.7ms JS)
    W->>G: Prepend new JSON entry to embedded data block
    W->>G: Insert new rows in summary table (native + wasm)
    W->>G: Update "Latest results" section to v3.3.0
    W->>PR: Open docs PR with updated INCREMENTAL-BENCHMARKS.md
Loading

Last reviewed commit: "docs: update increme..."

@carlos-alm carlos-alm merged commit d675ac1 into main Mar 19, 2026
14 checks passed
@carlos-alm carlos-alm deleted the benchmark/incremental-v3.3.0-20260319-081441 branch March 19, 2026 09:30
@github-actions github-actions bot locked and limited conversation to collaborators Mar 19, 2026
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant