[not-matthias]

Walltime mode now passes the same seed flags as analysis mode: --hash-seed=1 and --random-seed=1, plus --no-randomize-hashes on Node.js < 18. Both modes get them from one shared helper. Disabled mode stays unseeded, so runs outside CodSpeed keep the normal behavior.

Without these flags, every walltime process used a random Math.random() seed and a random hash seed. Benchmarks that use random inputs, or that depend on hash table layout (collisions, lookup cost), could vary between runs for reasons that have nothing to do with the code under test.

Checked on Node.js 24:

• Math.random() returns the same sequence on every run, including in worker threads and in vitest walltime forks.
• --hash-seed still applies even though Node boots from a snapshot, because --rehash-snapshot is on by default. With seed 1 the internal dictionary layout of the same object is identical across runs. Seed 2 gives a different but stable layout, and no flag gives a different layout each run.
• crypto randomness and key iteration order are unchanged.

--predictable and the GC flags stay analysis-only, because they change walltime timing.

The runner's exec-harness node wrapper copies these flags by hand and needs the same walltime update. Until then, checkV8Flags will warn about the missing seed flags under codspeed exec.

Stacked on #104.

Refs COD-3936

[codspeed]

Merging this PR will degrade performance by 1.13%

3 warnings, 1 note

⚠️ Walltime results may be unreliable

Some benchmarks used the Walltime instrument outside CodSpeed Macro Runners, which can make the results inconsistent.

Learn more about Macro Runners

See affected benchmarks

⚠️ 13 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

⚠️ Base and head ran on different runner types

Some benchmarks with significant performance changes were compared across different runner types, so the changes might not come from the code. Run base and head on the same runner type, or use CodSpeed Macro Runners for more stable results.

Learn more about Macro Runners

Open the report in CodSpeed to investigate

ℹ️ 16 benchmarks have no flamegraph

The flamegraph wasn't generated: most of the time was spent in runtime internals, which would leave nothing useful to analyze.

See affected benchmarks

⚡ 4 improved benchmarks
❌ 6 regressed benchmarks
✅ 254 untouched benchmarks
⏩ 56 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	CPU Simulation	wait 500ms	27.1 ms	48.4 ms	-44%
❌	CPU Simulation	wait 1sec	54.2 ms	91.3 ms	-40.65%
❌	CPU Simulation	test_recursive_fibo_10	339 µs	513.2 µs	-33.94%
❌	CPU Simulation	wait 500ms	27.1 ms	31.8 ms	-14.83%
❌	CPU Simulation	wait 1sec	55.7 ms	64.4 ms	-13.54%
❌	Walltime	switch 2	84 ns	96 ns	-12.5%
⚡	CPU Simulation	switch 1	657.1 µs	310.2 µs	×2.1
⚡	CPU Simulation	recursive fibo 10	685.3 µs	339.9 µs	×2
⚡	Memory	wait 500ms	28.7 KB	22.7 KB	+26.54%
⚡	Walltime	switch 1	84 ns	72 ns	+16.67%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing cod-3936-walltime-random-seed (c40ae87) with cod-3936-add-always-sparkplug-to-codspeed-node-analysis-mode-v8-flags (8dcd85b)}

Footnotes

1. 56 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩