CLI flag to enable tracing for benchmarks

[gatesn]

Similar to the --samply flag, the --tracing flag captures spans and log events for all threads.

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will degrade performance by 98.62%

_{Comparing ngates/perfetto (383d4ae) with develop (6b8fb88)¹}

Summary

⚡ 24 improved benchmarks
❌ 26 regressed benchmarks
✅ 1088 untouched benchmarks
⏩ 1265 skipped benchmarks²

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	take_fsl_random[64, 100]	152.6 µs	83.4 µs	+82.86%
❌	Simulation	bench_many_nulls[0.9]	469.4 µs	535.1 µs	-12.28%
⚡	Simulation	take_fsl_nullable_random[1024, 1000]	22.6 ms	8.9 ms	×2.5
⚡	Simulation	take_fsl_nullable_random[16, 100]	123.2 µs	109.4 µs	+12.7%
⚡	Simulation	take_fsl_nullable_random[256, 1000]	4.7 ms	2.2 ms	×2.2
⚡	Simulation	take_fsl_nullable_random[256, 100]	526 µs	261.4 µs	×2
⚡	Simulation	take_fsl_nullable_random[1024, 100]	1,987.4 µs	929.1 µs	×2.1
⚡	Simulation	take_fsl_nullable_random[4096, 1000]	90.8 ms	35.3 ms	×2.6
⚡	Simulation	take_fsl_nullable_random[4096, 100]	8.1 ms	3.6 ms	×2.3
⚡	Simulation	take_fsl_nullable_random[16, 1000]	399.7 µs	219.9 µs	+81.73%
⚡	Simulation	take_fsl_nullable_random[64, 100]	166.1 µs	100 µs	+66.06%
⚡	Simulation	take_fsl_nullable_random[64, 1000]	1,265.9 µs	609.7 µs	×2.1
⚡	Simulation	take_fsl_random[16, 100]	60.6 µs	42.2 µs	+43.6%
⚡	Simulation	take_fsl_random[16, 1000]	367.4 µs	178.6 µs	×2.1
⚡	Simulation	take_fsl_random[4096, 100]	8.2 ms	3.6 ms	×2.3
⚡	Simulation	take_fsl_random[1024, 100]	1,992.8 µs	913 µs	×2.2
⚡	Simulation	take_fsl_random[1024, 1000]	23 ms	8.9 ms	×2.6
⚡	Simulation	take_fsl_random[256, 100]	520.5 µs	249 µs	×2.1
⚡	Simulation	take_fsl_random[256, 1000]	4.9 ms	2.3 ms	×2.2
⚡	Simulation	take_fsl_random[4096, 1000]	92.7 ms	35.5 ms	×2.6
...	...	...	...	...	...

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

Footnotes

1. No successful run was found on develop (a11336d) during the generation of this report, so 6b8fb88 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

2. 1265 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. ↩