docs: update query benchmarks (3.9.6)#1031
Conversation
Greptile SummaryThis automated PR appends query benchmark results for version 3.9.6 to the benchmark history table and updates the "Latest results" section. The numeric data is internally consistent — all percentage delimiters ( Two minor points worth confirming: the Confidence Score: 4/5Safe to merge — documentation-only automated update with no logic changes and internally consistent data. Only P2 style/clarification comments exist. The data is self-consistent; the mid-target change and missing 3.9.5 row are questions rather than errors. No files require special attention beyond the clarification questions noted inline. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Workflow Run #786\nRelease 3.9.6] --> B[Collect query benchmarks\nMedian over 5 runs]
B --> C{Engine}
C --> D[Native / Rust\nhub=buildGraph, mid=dir, leaf=docs]
C --> E[WASM\nhub=buildGraph, mid=dir, leaf=docs]
D --> F[fnDeps d1/d3/d5\nfnImpact d1/d3/d5\ndiffImpact]
E --> F
F --> G[Append row to\nhistory table]
F --> H[Update 'Latest results'\ndetail section]
F --> I[Append JSON block\nfor machine consumption]
Reviews (1): Last reviewed commit: "docs: update query benchmarks (3.9.6)" | Re-trigger Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| | 3.9.6 | native | 28.7 ↑7% | 29.1 ↑7% | 33.2 ↑22% | 5.3 ↑8% | 5.5 ↑8% | 5.4 ↑6% | 16.7ms ↑129% | | ||
| | 3.9.6 | wasm | 38.4 ~ | 40 ↑4% | 38.4 ↓2% | 5 ~ | 5 ↓19% | 4.9 ↓20% | 11.5ms ↓20% | |
There was a problem hiding this comment.
The mid target node changed from ctx (3.9.4) to dir (3.9.6) in both native and WASM runs. The fnImpact depth metrics in the summary table reflect traversal from the hub (buildGraph), so they remain comparable, but benchmark runs that specifically exercise the mid-connectivity node are now on a different function. If ctx was renamed or removed in 3.9.6 this is expected — a small annotation in the "Latest results" section noting the target change would help readers interpret the relative ↑/↓ indicators for those runs. Was ctx renamed to dir in 3.9.6, or did the most representative mid-connectivity node simply shift? If the latter, are comparisons between 3.9.4 and 3.9.6 still meaningful for the mid-target benchmarks?
|
|
||
| | Version | Engine | fnDeps d1 | fnDeps d3 | fnDeps d5 | fnImpact d1 | fnImpact d3 | fnImpact d5 | diffImpact | | ||
| |---------|--------|----------:|----------:|----------:|------------:|------------:|------------:|-----------:| | ||
| | 3.9.6 | native | 28.7 ↑7% | 29.1 ↑7% | 33.2 ↑22% | 5.3 ↑8% | 5.5 ↑8% | 5.4 ↑6% | 16.7ms ↑129% | |
There was a problem hiding this comment.
The table jumps from 3.9.4 directly to 3.9.6 — there is no 3.9.5 row. If 3.9.5 was released but benchmarks weren't collected (e.g., it was a hotfix), this is fine, but worth confirming so readers don't assume it was not released. Was 3.9.5 released? If so, was it intentionally excluded from benchmarks (e.g., patch-only release)?
1 participant
Automated query benchmark update for 3.9.6 from workflow run #786.