Execute tensor product functions with RowFn

[connortsui20]

Summary

• Tracking Issue: Epic: Row-oriented scalar functions #9128
• Depends on: Execute tensor L2 norm with RowFn #9768

Moves inner product and cosine similarity to RowFn without changing their arithmetic contracts.

Changes

Inner product keeps the existing left-to-right multiply-and-sum order. Cosine computes that dot product followed by both L2 norms and the existing zero-denominator guard, with no encoded reduction, reassociation, or constant-norm shortcut. Tests cover nulls, empty and zero-width rows, IEEE overflow and underflow, and bitwise agreement between constant encodings and materialized rows.

[codspeed-hq]

Merging this PR will regress 25 benchmarks

⚠️ 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

⚡ 12 improved benchmarks
❌ 25 regressed benchmarks
✅ 2150 untouched benchmarks
⏩ 218 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	compress_alp[f64, (1000, 0.0, 1.0)]	105 µs	159 µs	-33.94%
❌	Simulation	compress_alp[f64, (1000, 0.0, 0.25)]	114.4 µs	168.2 µs	-32%
❌	Simulation	compress_alp[f64, (1000, 0.0, 0.95)]	114.2 µs	167.2 µs	-31.68%
❌	Simulation	compress_alp[f64, (1000, 0.1, 1.0)]	150.3 µs	211.9 µs	-29.09%
❌	Simulation	compress_alp[f64, (1000, 0.01, 1.0)]	144.2 µs	201.1 µs	-28.27%
❌	Simulation	alp_compress_f64	171.8 µs	237.8 µs	-27.75%
❌	Simulation	compress_alp[f64, (1000, 0.01, 0.25)]	155.1 µs	211.9 µs	-26.78%
❌	Simulation	compress_alp[f64, (1000, 0.1, 0.25)]	168.4 µs	228.9 µs	-26.46%
❌	Simulation	compress_alp[f64, (1000, 0.1, 0.95)]	177.5 µs	237.3 µs	-25.21%
❌	Simulation	compress_alp[f64, (1000, 0.01, 0.95)]	159.6 µs	213.2 µs	-25.13%
❌	Simulation	compress_alp[f32, (1000, 0.0, 1.0)]	51.9 µs	63.4 µs	-18.08%
❌	Simulation	compress_alp[f32, (1000, 0.0, 0.95)]	61.1 µs	73.1 µs	-16.47%
❌	Simulation	compress_alp[f32, (1000, 0.0, 0.25)]	61.2 µs	73.3 µs	-16.47%
❌	Simulation	compress_alp[f64, (10000, 0.0, 1.0)]	308.2 µs	361 µs	-14.64%
❌	Simulation	compress_alp[f64, (10000, 0.0, 0.95)]	318.3 µs	370.9 µs	-14.19%
❌	Simulation	compress_alp[f64, (10000, 0.0, 0.25)]	318.2 µs	370.7 µs	-14.16%
❌	Simulation	compress_alp[f64, (10000, 0.01, 1.0)]	538.8 µs	616.2 µs	-12.57%
❌	Simulation	compress_alp[f32, (1000, 0.1, 1.0)]	96.1 µs	109.5 µs	-12.19%
❌	Simulation	compress_alp[f32, (1000, 0.01, 1.0)]	90.6 µs	103.1 µs	-12.08%
❌	Simulation	compress_alp[f64, (10000, 0.1, 1.0)]	574 µs	652.6 µs	-12.05%
...	...	...	...	...	...

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 ct/row-fn-tensor-products-v2 (b37f665) with ct/row-fn-tensor-l2-v2 (5974184)²}

Footnotes

1. 218 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. No successful run was found on ct/row-fn-tensor-l2-v2 (929ae6d) during the generation of this report, so da9d39c was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩