fix: [profile.bench] codegen-units = 16

[joseph-isaacs]

#7824 changed this field?

We use 16 since this is cargo release default.

[AdamGS]

don't we want to bench in the most optimized profile?

[AdamGS]

having a description/explaination in the commit message would be useful here, or maybe a comment in the file itself.

[joseph-isaacs]

We might want that but we should change this in a stand alone PR

[AdamGS]

why? if we're going to revert this we can just directly get to the end state

[codspeed-hq]

Merging this PR will degrade performance by 16.24%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 12 improved benchmarks
❌ 147 regressed benchmarks
✅ 1348 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	from_iter_bit_buffer[65536]	113.4 µs	241.3 µs	-52.99%
❌	Simulation	from_iter_bit_buffer[16384]	32.4 µs	64.7 µs	-49.92%
❌	Simulation	from_iter_bit_buffer[2048]	7.8 µs	12.2 µs	-36.03%
❌	Simulation	chunked_varbinview_canonical_into[(100, 100)]	269.9 µs	398.9 µs	-32.33%
❌	Simulation	chunked_varbinview_canonical_into[(1000, 10)]	197.1 µs	284.6 µs	-30.76%
❌	Simulation	binary_search_vortex	486.1 ns	698.6 ns	-30.42%
❌	Simulation	chunked_varbinview_into_canonical[(100, 100)]	320.1 µs	457.3 µs	-29.99%
❌	Simulation	chunked_varbinview_into_canonical[(1000, 10)]	211.7 µs	300.7 µs	-29.59%
❌	Simulation	from_iter_bit_buffer[1024]	6.2 µs	8.6 µs	-28.12%
❌	Simulation	chunked_varbinview_opt_canonical_into[(1000, 10)]	223.9 µs	308.4 µs	-27.4%
❌	Simulation	compare[62]	316.8 µs	432.3 µs	-26.71%
❌	Simulation	compare[63]	322.5 µs	439.8 µs	-26.68%
❌	Simulation	compare[60]	308.5 µs	420.3 µs	-26.59%
❌	Simulation	compare[56]	292.6 µs	396.8 µs	-26.27%
❌	Simulation	chunked_varbinview_opt_into_canonical[(1000, 10)]	239 µs	323.9 µs	-26.23%
❌	Simulation	compare[61]	318.2 µs	431.3 µs	-26.22%
❌	Simulation	compare[58]	307.8 µs	415.7 µs	-25.96%
❌	Simulation	compare[59]	313 µs	422.8 µs	-25.96%
❌	Simulation	varbinview_large	130.5 µs	175.4 µs	-25.6%
❌	Simulation	compare[57]	308.9 µs	414.6 µs	-25.49%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

Tip

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

---

_{Comparing ji/bench-fix (a248249) with develop (583b003)}

[joseph-isaacs]

Going to merge to speed up the CI.

I think it might be worth looking at the diff and add some inlines?