[0ax1]

Problem

Closing a mapped reader can stall decoding while waiting for GPU copies and releasing page tables.

Solution

Move synchronization and mapping release to a background worker; drain queued releases before explicit cache eviction.

Assessment

Falls back to synchronous release if queuing fails and retains mappings if completion is unknown. Depends on #10717.

[codspeed]

Merging this PR will degrade performance by 12.09%

2 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

⚠️ 21 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

ℹ️ 40 benchmarks have no flamegraph

The flamegraph wasn't generated: the symbols of the profiling trace could not be resolved, because no information about the language frames was collected.

See affected benchmarks

⚡ 7 improved benchmarks
❌ 8 regressed benchmarks
✅ 2393 untouched benchmarks
⏩ 12 skipped benchmarks¹
🗄️ 43 archived benchmarks run²

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Walltime	threshold_neon[8]	1.4 µs	3.1 µs	-55.7%
❌	Walltime	filter_neon[8]	932 ns	1,788 ns	-47.87%
❌	Walltime	threshold_neon[64]	2.6 µs	4.2 µs	-38.41%
❌	Walltime	filter_neon[64]	1.8 µs	2.8 µs	-34.96%
❌	Walltime	filter_neon[80]	2.7 µs	3.4 µs	-22.37%
❌	Walltime	threshold_neon[80]	3.8 µs	4.5 µs	-15.67%
❌	CPU Simulation	test_repeated_scan_scalar_at[int_str_float]	493.5 µs	577.2 µs	-14.51%
❌	CPU Simulation	test_repeated_scan_scalar_at[int_float]	373.4 µs	423.3 µs	-11.79%
⚡	CPU Simulation	decode_primitives[i64, (1000, 8)]	37.6 µs	24.2 µs	+55.51%
⚡	CPU Simulation	test_scan_filter_to_arrow_array[int_float]	4.1 ms	3.5 ms	+15.98%
⚡	Walltime	starts_with_avx512[8, VarBin]	46.2 µs	39.9 µs	+15.87%
⚡	Walltime	starts_with_avx512[64, VarBin]	62.8 µs	56.4 µs	+11.51%
⚡	CPU Simulation	test_repeated_scan_filter[int]	2.7 ms	2.5 ms	+11.39%
⚡	CPU Simulation	test_scan_filter_to_arrow_array[int]	2.8 ms	2.5 ms	+11.07%
⚡	CPU Simulation	test_scan_filter[int]	3 ms	2.8 ms	+10.02%

Tip

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

---

_{Comparing ad/cuda-mapped-release (8845ef7) with ad/cuda-file-mapping-cache (a149621)}

Footnotes

1. 12 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. ↩

2. 43 benchmarks were run, but are now archived. If they were deleted in another branch, consider rebasing to remove them from the report. Instead if they were added back, click here to restore them. ↩

[robert3005]

I wonder if there's a thing that already runs on a per thread basis that we can move this to instead of having a thread