GPU benchmark: add low-overhead Nsight ranges

[joseph-isaacs]

Stacked only on #9432, whose root is current develop; it has no dependency on the Parquet benchmark PR.

What

• Add VORTEX_GPU_PROFILE=nsys to wrap each field decode in an NVTX range while retaining encoding-tree metrics.
• Add VORTEX_GPU_PROFILE=nsys-ranges for lower-overhead captures that skip encoding-tree construction.

Why this is needed

Nsight Systems shows CUDA API calls, copies, waits, allocations, and kernel gaps, but those events are hard to map back to Vortex fields. NVTX ranges provide that correlation. The ranges-only mode is separate because formatting every full array/encoding tree materially perturbs the host timeline we are trying to inspect.

Validation

• cargo +nightly fmt --all
• cargo check -p compress-bench --features cuda,unstable_encodings
• git diff --check

[codspeed-hq]

Merging this PR will regress 1 benchmark

⚠️ 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.

⚡ 2 improved benchmarks
❌ 1 regressed benchmark
✅ 1997 untouched benchmarks
⏩ 89 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	compress_fsst[(1000, 64, 8)]	1 ms	1.1 ms	-10.76%
⚡	Simulation	decompress[u64, (10000, 4)]	402.6 µs	310.6 µs	+29.6%
⚡	Simulation	take[core/shuffled/primitive/nonnull/chunks=1024/indices=10000]	14.2 ms	12.9 ms	+10.18%

Tip

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

---

_{Comparing joe/gpu-nsys-range-profiling (d7c32df) with joe/gpu-decompression-profiling (b6cf956)}

Footnotes

1. 89 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. ↩