Support creating accumulators with already derived dtypes

[robert3005]

Avoid having to rederive accumulator dtypes for every group when combining them

[codspeed]

Merging this PR will degrade performance by 6.29%

⚠️ 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.

⚡ 3 improved benchmarks
❌ 7 regressed benchmarks
✅ 2198 untouched benchmarks
⏩ 293 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	decode_primitives[f32, (1000, 512)]	42 µs	64.3 µs	-34.74%
❌	WallTime	dbp_assemble_kernel_avx2[(I128, 1024)]	464 ns	569 ns	-18.45%
❌	WallTime	dbp_assemble_kernel_avx512[(I128, 1024)]	463 ns	549 ns	-15.66%
❌	Simulation	take_fsl_u32_random[64, 100]	158.8 µs	187.2 µs	-15.17%
❌	WallTime	filtered_owned_i64_avx2[OneNullInEight]	22.6 µs	25.8 µs	-12.46%
❌	Simulation	take_fsl_f16_random[256, 100]	204.9 µs	231.1 µs	-11.32%
❌	WallTime	dbp_assemble_kernel_narrow_msp_avx2[(I128, 1024)]	629 ns	705 ns	-10.78%
⚡	Simulation	take_fsl_f16_random[16, 100]	179.5 µs	120 µs	+49.61%
⚡	WallTime	filtered_sink_i64_avx2[OneNullInEight]	26.2 µs	22.3 µs	+17.09%
⚡	Simulation	take_fsl_nullable_random[16, 100]	188 µs	166.3 µs	+13.04%

Tip

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

---

_{Comparing rk/accumcombined (fe06585) with develop (a542cbd)}

Footnotes

1. 293 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. ↩