| name | run-bench |
| description | Execute the benchmark harness under benchmarks/ and produce a markdown
report comparing this build against the recorded baseline (and optionally
against Luerl and PUC-Lua). Flag regressions over a configurable threshold.
Use when the user asks for a benchmark run, after any executor or codegen
change in Direction B, or when a Direction B plan's verification calls
for it.
STATUS: stub. Will be filled in when Direction B begins. The benchee
harness is already in benchmarks/ (added in PR #143).
|
run-bench (stub)
This skill is a placeholder. It will be fleshed out when Direction B starts.
What's already in place (don't re-build)
benchmarks/ directory with benchee scripts: closures.exs, fibonacci.exs,
oop.exs, string_ops.exs, table_ops.exs. (PR #143)
- Comparison against Luerl and PUC-Lua via
luerl and luaport deps.
What this skill needs to add when Direction B starts
- A standard "run all benchmarks" entry point (probably a mix task or shell
script).
- Output to
bench/results/<date>-<sha>.md so we have a history.
- A baseline stored at
bench/baseline.md that gets updated explicitly
(not on every run).
- A diff renderer: takes two result sets, produces a markdown table
showing percent change, flags regressions over a threshold (default 20%).
- Convention for capturing PR-relevant numbers in the PR body.
Until then
Run benchmarks manually:
mix run benchmarks/fibonacci.exs
mix run benchmarks/closures.exs
mix run benchmarks/oop.exs
mix run benchmarks/string_ops.exs
mix run benchmarks/table_ops.exs
Capture stdout and paste into the PR description. Until a baseline file
exists, "no regression" is judged by re-running before and after on the
same machine.