perf: non-blocking Arrow Device array release

[0ax1]

No description provided.

[codspeed-hq]

Merging this PR will improve performance by 24.25%

⚠️ 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.

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 111 improved benchmarks
❌ 5 regressed benchmarks
✅ 1412 untouched benchmarks
⏩ 10 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	compare[12]	115.8 µs	136.4 µs	-15.08%
❌	WallTime	cuda/bitpacked_u8/unpack/3bw[100M]	299.1 µs	352.1 µs	-15.04%
❌	Simulation	varbinview_large	112.9 µs	131.5 µs	-14.19%
❌	Simulation	compare[8]	105 µs	118 µs	-11.05%
❌	Simulation	compare[10]	135.9 µs	152.7 µs	-11.02%
⚡	Simulation	compare[48]	300.6 µs	213 µs	+41.15%
⚡	Simulation	compare[50]	319.2 µs	227.7 µs	+40.18%
⚡	Simulation	compare[49]	317.7 µs	228.2 µs	+39.24%
⚡	Simulation	compare[44]	287.7 µs	207.5 µs	+38.68%
⚡	Simulation	compare[46]	302.5 µs	218.5 µs	+38.46%
⚡	Simulation	compare[47]	309.4 µs	223.5 µs	+38.4%
⚡	Simulation	compare[40]	263.5 µs	190.7 µs	+38.18%
⚡	Simulation	compare[44]	292.4 µs	212.1 µs	+37.82%
⚡	Simulation	compare[45]	301 µs	218.9 µs	+37.49%
⚡	Simulation	compare[43]	287.6 µs	209.2 µs	+37.47%
⚡	Simulation	compare[42]	281.1 µs	204.5 µs	+37.42%
⚡	Simulation	compare[40]	268.3 µs	195.6 µs	+37.17%
⚡	Simulation	compare[43]	292.5 µs	214.2 µs	+36.57%
⚡	Simulation	compare[42]	286 µs	209.4 µs	+36.56%
⚡	Simulation	compare[41]	279.2 µs	204.5 µs	+36.49%
...	...	...	...	...	...

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 ad/arrow-device-release-ordering (b1e780d) with develop (e74aac3)}

Footnotes

1. 10 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. ↩