Expose CUDA-compatible file sink through FFI

[onursatici]

Rationale for this change

cuda compatible vortex files currently can't be written over ffi

What changes are included in this PR?

ffi to allow writing with the cuda compatible strategy, which only includes encodings that we have gpu kernels for, and also uses the cuda flat layout that uses the array tree from the footer. Also exclude alp-rd and pco from the cuda compatible encoding allowlist for completeness

[github-actions]

Benchmarks: Vortex queries

Verdict: Likely regression (high confidence)
Attributed Vortex impact: +12.3%
Engines: DataFusion No clear signal (+7.8%, low confidence) · DuckDB Likely regression (+17.0%, high confidence)
Vortex (geomean): 1.166x ❌
Parquet (geomean): 1.100x ❌
Shifts: Parquet (control) +10.0% · Median polish +18.0%

How to read Verdict and Engines

• Verdict: Overall PR-level signal after subtracting baseline drift estimated from Parquet control rows. It can be Likely improvement, Likely regression, or No clear signal.
• Engines: Per-engine attribution. DataFusion is compared against DataFusion/Parquet controls; DuckDB is compared against DuckDB/Parquet controls. This answers whether each engine improved or regressed independently.
• Confidence: Based on directional consistency, share of rows above the noise floor, and control-run noise.

---

datafusion / vortex-file-compressed (1.225x ❌, 0↑ 2↓)

name	PR 9a15dd8 (ns)	base 1c419e5 (ns)	ratio (PR/base)
vortex_q00/datafusion:vortex-file-compressed 🚨	11775292	9704767	1.21
vortex_q01/datafusion:vortex-file-compressed 🚨	7441524	6015063	1.24

datafusion / parquet (1.136x ❌, 0↑ 2↓)

name	PR 9a15dd8 (ns)	base 1c419e5 (ns)	ratio (PR/base)
vortex_q00/datafusion:parquet 🚨	23466916	20806650	1.13
vortex_q01/datafusion:parquet 🚨	5416210	4731642	1.14

duckdb / vortex-file-compressed (1.247x ❌, 0↑ 2↓)

name	PR 9a15dd8 (ns)	base 1c419e5 (ns)	ratio (PR/base)
vortex_q00/duckdb:vortex-file-compressed 🚨	12521427	10153450	1.23
vortex_q01/duckdb:vortex-file-compressed 🚨	7278018	5772282	1.26

duckdb / parquet (1.065x ➖, 0↑ 0↓)

name	PR 9a15dd8 (ns)	base 1c419e5 (ns)	ratio (PR/base)
vortex_q00/duckdb:parquet	24914677	23232389	1.07
vortex_q01/duckdb:parquet	9916134	9368910	1.06

---

No file size changes detected.

[myrrc]

LGTM for vortex-ffi changes

[github-actions]

Polar Signals Profiling Results

Latest Run

Status	Commit	Job	Attempt	Link
🟢 Done	9a15dd8		1	Explore Profiling Data

---

Powered by Polar Signals Cloud

[codspeed-hq]

Merging this PR will degrade performance by 13.22%

⚠️ 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.

❌ 1 regressed benchmark
✅ 1697 untouched benchmarks
⏩ 9 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	WallTime	cuda/bitpacked_u8/unpack/3bw[100M]	301.6 µs	347.5 µs	-13.22%

Tip

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

---

_{Comparing os/ffi-cuda (9a15dd8) with develop (1c419e5)}

Footnotes

1. 9 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. ↩