[connortsui20]

Summary

vx.scalar(decimal.Decimal("9999999999999999999")) fails with an overflow error, and so does a JNI decimal literal at precision 19 with a value above i64::MAX. Both factories picked the value width from DecimalDType::required_bit_width(), which counts magnitude bits without a sign bit. At precision 19 that is 64, so they chose i64, which can't hold the largest 19-digit values.

Changes

Both factories now match on DecimalDType::canonical_values_type() from #10897, which gives i128 at precision 19. Every other precision keeps its width, because the sign bit still fits there. The Python precision inference test and a new JNI unit test cover the largest and smallest values at precision 19.

This is stacked on #10897 and retargets to develop after that merges.

[connortsui20]

This might blow things up but I think that is fine because it is a scalar, not an array?

[codspeed]

Merging this PR will degrade performance by 13.54%

3 warnings

⚠️ 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

⚠️ 3 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

⚠️ Base and head ran on different runner types

Some benchmarks with significant performance changes were compared across different runner types, so the changes might not come from the code. Run base and head on the same runner type, or use CodSpeed Macro Runners for more stable results.

Learn more about Macro Runners

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
❌ 3 regressed benchmarks
✅ 2200 untouched benchmarks
⏩ 370 skipped benchmarks¹
🗄️ 30 archived benchmarks run²

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Walltime	threshold_neon[64]	2.6 µs	3.8 µs	-31.52%
❌	Walltime	threshold_neon[80]	3.3 µs	4.4 µs	-24.01%
❌	Walltime	filter_neon[80]	2.4 µs	2.8 µs	-12.32%
⚡	CPU Simulation	test_repeated_scan_scalar_at[int_str]	412.1 µs	336.5 µs	+22.46%

Tip

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

---

_{Comparing ct/decimal-scalar-factory-width (41073b9) with ct/decimal-width-owner (9836fef)}

Footnotes

1. 370 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. 30 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. ↩