Latest known benchmark result, per method

Generated. Do not edit. Produced by .github/workflows/bench-nightly.yml from the wiki's own history, alongside nightly_run.

Not a comparison across methods. Each section below is the last night that method was actually re-run -- whichever night touched the source near it, not necessarily last night, and not the same night as its neighbours here. Every run measures on a GitHub hosted runner whose hardware differs night to night, so a number here says "this is the last known reading", never "faster than the section above it".

Per method

Lodestar.Gpu.Benchmarks.BitParallelEditDistanceBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Texts	TextLength	Device	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
CpuBaseline	10000	32	CPUAc(...) GPU) [26]	2.206 ms	0.1401 ms	0.0077 ms	1.00	0.00	-	-	-	-	NA
GpuResident	10000	32	CPUAc(...) GPU) [26]	4.417 ms	8.6152 ms	0.4722 ms	2.00	0.19	-	-	-	42832 B	NA
GpuFromHost	10000	32	CPUAc(...) GPU) [26]	5.191 ms	7.6277 ms	0.4181 ms	2.35	0.16	93.7500	93.7500	93.7500	471051 B	NA
CpuBaseline	10000	256	CPUAc(...) GPU) [26]	10.702 ms	0.0452 ms	0.0025 ms	1.00	0.00	-	-	-	-	NA
GpuResident	10000	256	CPUAc(...) GPU) [26]	6.743 ms	7.4163 ms	0.4065 ms	0.63	0.03	-	-	-	42840 B	NA
GpuFromHost	10000	256	CPUAc(...) GPU) [26]	9.600 ms	6.6430 ms	0.3641 ms	0.90	0.03	328.1250	328.1250	328.1250	2711059 B	NA
CpuBaseline	200000	32	CPUAc(...) GPU) [26]	45.753 ms	0.7169 ms	0.0393 ms	1.00	0.00	-	-	-	-	NA
GpuResident	200000	32	CPUAc(...) GPU) [26]	36.352 ms	40.6641 ms	2.2289 ms	0.79	0.04	-	-	-	802909 B	NA
GpuFromHost	200000	32	CPUAc(...) GPU) [26]	43.333 ms	29.6994 ms	1.6279 ms	0.95	0.03	454.5455	454.5455	454.5455	8071853 B	NA
CpuBaseline	200000	256	CPUAc(...) GPU) [26]	214.959 ms	0.2871 ms	0.0157 ms	1.00	0.00	-	-	-	-	NA
GpuResident	200000	256	CPUAc(...) GPU) [26]	119.582 ms	16.6947 ms	0.9151 ms	0.56	0.00	-	-	-	803029 B	NA
GpuFromHost	200000	256	CPUAc(...) GPU) [26]	193.391 ms	65.4286 ms	3.5864 ms	0.90	0.01	333.3333	333.3333	333.3333	52871293 B	NA

Lodestar.Gpu.Benchmarks.ChainedProductBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Width	Device	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
RoundTripped	2000	32	CPUAc(...) GPU) [26]	1,439.671 ms	183.1393 ms	10.0385 ms	1.000	0.01	-	-	-	7.32 MB	1.00
Chained	2000	32	CPUAc(...) GPU) [26]	1,594.555 ms	2,073.4169 ms	113.6510 ms	1.108	0.07	-	-	-	5.36 MB	0.73
CpuBaseline	2000	32	CPUAc(...) GPU) [26]	1.140 ms	0.0579 ms	0.0032 ms	0.001	0.00	308.5938	308.5938	308.5938	1.47 MB	0.20
RoundTripped	2000	128	CPUAc(...) GPU) [26]	6,068.606 ms	4,562.5821 ms	250.0905 ms	1.001	0.05	1000.0000	1000.0000	1000.0000	37.82 MB	1.00
Chained	2000	128	CPUAc(...) GPU) [26]	6,220.553 ms	1,602.5821 ms	87.8429 ms	1.026	0.04	1000.0000	1000.0000	1000.0000	35.45 MB	0.94
CpuBaseline	2000	128	CPUAc(...) GPU) [26]	5.128 ms	0.6662 ms	0.0365 ms	0.001	0.00	492.1875	492.1875	492.1875	5.86 MB	0.15
RoundTripped	20000	32	CPUAc(...) GPU) [26]	6,065.032 ms	1,688.9975 ms	92.5797 ms	1.000	0.02	1000.0000	1000.0000	1000.0000	44.09 MB	1.00
Chained	20000	32	CPUAc(...) GPU) [26]	6,111.526 ms	1,540.3258 ms	84.4305 ms	1.008	0.02	1000.0000	1000.0000	1000.0000	42.21 MB	0.96
CpuBaseline	20000	32	CPUAc(...) GPU) [26]	6.203 ms	2.0169 ms	0.1106 ms	0.001	0.00	390.6250	390.6250	390.6250	5.86 MB	0.13
RoundTripped	20000	128	CPUAc(...) GPU) [26]	25,113.270 ms	2,402.8009 ms	131.7056 ms	1.000	0.01	15000.0000	14000.0000	13000.0000	388.64 MB	1.00
Chained	20000	128	CPUAc(...) GPU) [26]	25,186.220 ms	3,449.3382 ms	189.0699 ms	1.003	0.01	20000.0000	19000.0000	18000.0000	384.96 MB	0.99
CpuBaseline	20000	128	CPUAc(...) GPU) [26]	12.416 ms	5.9356 ms	0.3254 ms	0.000	0.00	484.3750	484.3750	484.3750	23.44 MB	0.06

Lodestar.Gpu.Benchmarks.CosineTopKSelectionBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	K	Device	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
SimdBaseline	10000	10	CPUAc(...) GPU) [26]	7.007 ms	0.5991 ms	0.0328 ms	1.00	0.01	26.01 KB	1.00
GpuResident	10000	10	CPUAc(...) GPU) [26]	23,185.352 ms	3,317.0949 ms	181.8212 ms	3,309.16	26.17	1519.13 KB	58.41
SimdBaseline	10000	100	CPUAc(...) GPU) [26]	7.494 ms	0.0707 ms	0.0039 ms	1.00	0.00	37.26 KB	1.00
GpuResident	10000	100	CPUAc(...) GPU) [26]	24,372.652 ms	13,636.7450 ms	747.4760 ms	3,252.10	86.39	1541.63 KB	41.38
SimdBaseline	10000	1000	CPUAc(...) GPU) [26]	11.598 ms	0.1826 ms	0.0100 ms	1.00	0.00	149.77 KB	1.00
GpuResident	10000	1000	CPUAc(...) GPU) [26]	27,127.554 ms	1,756.8404 ms	96.2984 ms	2,338.93	7.40	1765.7 KB	11.79
SimdBaseline	100000	10	CPUAc(...) GPU) [26]	83.646 ms	47.5531 ms	2.6065 ms	1.00	0.04	26.14 KB	1.00
GpuResident	100000	10	CPUAc(...) GPU) [26]	227,672.594 ms	44,967.3860 ms	2,464.8142 ms	2,723.60	77.17	39277.98 KB	1,502.45
SimdBaseline	100000	100	CPUAc(...) GPU) [26]	77.132 ms	3.5019 ms	0.1920 ms	1.00	0.00	37.39 KB	1.00
GpuResident	100000	100	CPUAc(...) GPU) [26]	228,960.147 ms	35,223.1903 ms	1,930.7019 ms	2,968.44	22.60	39300.76 KB	1,051.03
SimdBaseline	100000	1000	CPUAc(...) GPU) [26]	83.152 ms	3.9236 ms	0.2151 ms	1.00	0.00	149.92 KB	1.00
GpuResident	100000	1000	CPUAc(...) GPU) [26]	233,960.983 ms	12,308.2118 ms	674.6546 ms	2,813.66	9.44	39526.7 KB	263.66

Lodestar.Gpu.Benchmarks.TiledMinHashSignaturesBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Permutations	Device	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
CpuBaseline	5000	64	CPUAc(...) GPU) [26]	98.61 ms	3.705 ms	0.203 ms	1.00	0.00	166.6667	-	-	2.98 MB	1.00
GpuResident	5000	64	CPUAc(...) GPU) [26]	4,737.31 ms	3,258.235 ms	178.595 ms	48.04	1.57	-	-	-	5.81 MB	1.95
GpuWithHashing	5000	64	CPUAc(...) GPU) [26]	5,010.71 ms	587.103 ms	32.181 ms	50.81	0.30	-	-	-	10.94 MB	3.68
CpuBaseline	5000	128	CPUAc(...) GPU) [26]	103.16 ms	4.485 ms	0.246 ms	1.00	0.00	200.0000	-	-	4.2 MB	1.00
GpuResident	5000	128	CPUAc(...) GPU) [26]	9,812.77 ms	5,548.253 ms	304.118 ms	95.12	2.56	-	-	-	13.78 MB	3.28
GpuWithHashing	5000	128	CPUAc(...) GPU) [26]	10,083.91 ms	6,615.448 ms	362.615 ms	97.75	3.05	-	-	-	18.91 MB	4.51
CpuBaseline	50000	64	CPUAc(...) GPU) [26]	998.84 ms	36.428 ms	1.997 ms	1.00	0.00	1000.0000	-	-	29.76 MB	1.00
GpuResident	50000	64	CPUAc(...) GPU) [26]	49,259.59 ms	10,516.422 ms	576.441 ms	49.32	0.51	4000.0000	3000.0000	2000.0000	160.03 MB	5.38
GpuWithHashing	50000	64	CPUAc(...) GPU) [26]	49,773.96 ms	11,530.669 ms	632.035 ms	49.83	0.55	7000.0000	4000.0000	2000.0000	211.59 MB	7.11
CpuBaseline	50000	128	CPUAc(...) GPU) [26]	1,020.81 ms	21.754 ms	1.192 ms	1.00	0.00	2000.0000	-	-	41.96 MB	1.00
GpuResident	50000	128	CPUAc(...) GPU) [26]	100,167.39 ms	15,148.210 ms	830.325 ms	98.13	0.71	12000.0000	10000.0000	8000.0000	536.13 MB	12.78
GpuWithHashing	50000	128	CPUAc(...) GPU) [26]	99,654.64 ms	8,276.622 ms	453.670 ms	97.62	0.40	15000.0000	12000.0000	8000.0000	587.65 MB	14.00

Lodestar.Gpu.Benchmarks.TiledSparseDenseProductBenchmarks-report-github

As of 2026-10-16, measured at commit 724c791262e7139428ab02fd4e70a6e05305bac7.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.402
  [Host]   : .NET 10.0.13 (10.0.1326.50504), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.13 (10.0.1326.50504), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Width	Device	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
CpuBaseline	5000	64	CPUAc(...) GPU) [26]	4.736 ms	0.1452 ms	0.0080 ms	1.00	0.00	148.4375	148.4375	148.4375	2.44 MB	1.00
GpuResident	5000	64	CPUAc(...) GPU) [26]	7,936.388 ms	10,448.8434 ms	572.7364 ms	1,675.93	104.77	-	-	-	20.95 MB	8.58
GpuFromHost	5000	64	CPUAc(...) GPU) [26]	7,583.749 ms	3,533.5361 ms	193.6850 ms	1,601.47	35.50	-	-	-	23.26 MB	9.53
CpuBaseline	5000	256	CPUAc(...) GPU) [26]	15.436 ms	2.7440 ms	0.1504 ms	1.00	0.01	125.0000	125.0000	125.0000	9.77 MB	1.00
GpuResident	5000	256	CPUAc(...) GPU) [26]	30,068.332 ms	6,222.5518 ms	341.0791 ms	1,948.06	25.18	1000.0000	-	-	139.74 MB	14.31
GpuFromHost	5000	256	CPUAc(...) GPU) [26]	29,886.177 ms	6,531.8107 ms	358.0306 ms	1,936.26	25.85	1000.0000	-	-	146.05 MB	14.96
CpuBaseline	50000	64	CPUAc(...) GPU) [26]	46.911 ms	1.4874 ms	0.0815 ms	1.00	0.00	181.8182	181.8182	181.8182	24.41 MB	1.00
GpuResident	50000	64	CPUAc(...) GPU) [26]	74,761.614 ms	15,795.4996 ms	865.8046 ms	1,593.68	16.16	10000.0000	9000.0000	8000.0000	519.77 MB	21.29
GpuFromHost	50000	64	CPUAc(...) GPU) [26]	75,025.414 ms	14,813.2074 ms	811.9619 ms	1,599.30	15.18	11000.0000	10000.0000	9000.0000	546.85 MB	22.40
CpuBaseline	50000	256	CPUAc(...) GPU) [26]	168.351 ms	79.0823 ms	4.3348 ms	1.00	0.03	250.0000	250.0000	250.0000	97.66 MB	1.00
GpuResident	50000	256	CPUAc(...) GPU) [26]	308,955.099 ms	21,708.5470 ms	1,189.9187 ms	1,836.01	42.01	83000.0000	82000.0000	72000.0000	7354.9 MB	75.31
GpuFromHost	50000	256	CPUAc(...) GPU) [26]	303,791.917 ms	16,859.3839 ms	924.1197 ms	1,805.33	41.14	81000.0000	80000.0000	70000.0000	7381.63 MB	75.59

Lodestar.Stats.Benchmarks.CorrelationBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	PairCount	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Pearson	100	994.5 ns	235.48 ns	12.91 ns	1.00	0.02	-	-	-	-	NA
MetaNumerics_Pearson	100	797.4 ns	13.07 ns	0.72 ns	0.80	0.01	0.0095	-	-	168 B	NA
Lodestar_Spearman	100	2,815.8 ns	537.24 ns	29.45 ns	2.83	0.04	0.2441	-	-	4144 B	NA
MetaNumerics_Spearman	100	3,764.4 ns	321.86 ns	17.64 ns	3.79	0.04	0.1183	-	-	2024 B	NA
Lodestar_KendallTau	100	5,322.2 ns	186.15 ns	10.20 ns	5.35	0.06	-	-	-	-	NA
MetaNumerics_KendallTau	100	10,686.2 ns	3,356.78 ns	184.00 ns	10.75	0.20	-	-	-	152 B	NA
Lodestar_KendallTau_Asymptotic	100	5,326.3 ns	625.35 ns	34.28 ns	5.36	0.07	-	-	-	-	NA
Lodestar_Pearson	10000	86,193.2 ns	4,220.36 ns	231.33 ns	1.00	0.00	-	-	-	-	NA
MetaNumerics_Pearson	10000	56,781.1 ns	1,940.50 ns	106.37 ns	0.66	0.00	-	-	-	168 B	NA
Lodestar_Spearman	10000	1,187,253.5 ns	51,602.40 ns	2,828.50 ns	13.77	0.04	166.0156	166.0156	166.0156	1324850 B	NA
MetaNumerics_Spearman	10000	1,947,312.8 ns	109,999.25 ns	6,029.43 ns	22.59	0.08	7.8125	-	-	160427 B	NA
Lodestar_KendallTau	10000	2,131,296.6 ns	143,763.56 ns	7,880.17 ns	24.73	0.10	-	-	-	-	NA
MetaNumerics_KendallTau	10000	290,929,022.8 ns	48,811,305.04 ns	2,675,512.35 ns	3,375.33	28.00	-	-	-	-	NA
Lodestar_KendallTau_Asymptotic	10000	2,125,222.2 ns	119,327.12 ns	6,540.72 ns	24.66	0.09	-	-	-	-	NA

Lodestar.Stats.Benchmarks.DistributionIncumbentBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar_StudentCdf	2,438.82 ns	318.644 ns	17.466 ns	1.000	0.01	-	-	NA
MathNet_StudentCdf	299.70 ns	0.486 ns	0.027 ns	0.123	0.00	-	-	NA
MetaNumerics_StudentCdf	342.07 ns	64.842 ns	3.554 ns	0.140	0.00	-	-	NA
Lodestar_StudentQuantile	7,395.61 ns	63.382 ns	3.474 ns	3.033	0.02	-	-	NA
MathNet_StudentQuantile	5,632.60 ns	278.939 ns	15.290 ns	2.310	0.02	-	112 B	NA
MetaNumerics_StudentQuantile	1,189.67 ns	97.357 ns	5.336 ns	0.488	0.00	0.0172	288 B	NA
Lodestar_FisherCdf	782.53 ns	12.052 ns	0.661 ns	0.321	0.00	-	-	NA
MathNet_FisherCdf	132.85 ns	9.888 ns	0.542 ns	0.054	0.00	-	-	NA
MetaNumerics_FisherCdf	139.73 ns	4.788 ns	0.262 ns	0.057	0.00	-	-	NA
Lodestar_FisherQuantile	2,840.45 ns	43.729 ns	2.397 ns	1.165	0.01	-	-	NA
MathNet_FisherQuantile	2,703.21 ns	437.575 ns	23.985 ns	1.108	0.01	0.0038	104 B	NA
MetaNumerics_FisherQuantile	320.51 ns	19.808 ns	1.086 ns	0.131	0.00	-	-	NA
Lodestar_ChiSquaredCdf	29.92 ns	0.241 ns	0.013 ns	0.012	0.00	-	-	NA
MathNet_ChiSquaredCdf	61.60 ns	6.437 ns	0.353 ns	0.025	0.00	-	-	NA
MetaNumerics_ChiSquaredCdf	75.83 ns	3.817 ns	0.209 ns	0.031	0.00	-	-	NA
Lodestar_ChiSquaredQuantile	367.96 ns	16.822 ns	0.922 ns	0.151	0.00	-	-	NA
MathNet_ChiSquaredQuantile	474.11 ns	5.462 ns	0.299 ns	0.194	0.00	-	-	NA
MetaNumerics_ChiSquaredQuantile	390.46 ns	7.371 ns	0.404 ns	0.160	0.00	-	-	NA
Lodestar_NormalCdf	21.78 ns	2.984 ns	0.164 ns	0.009	0.00	-	-	NA
MathNet_NormalCdf	22.31 ns	0.107 ns	0.006 ns	0.009	0.00	-	-	NA
MetaNumerics_NormalCdf	65.10 ns	0.135 ns	0.007 ns	0.027	0.00	-	-	NA
Lodestar_NormalQuantile	28.53 ns	0.477 ns	0.026 ns	0.012	0.00	-	-	NA
MathNet_NormalQuantile	30.50 ns	0.227 ns	0.012 ns	0.013	0.00	-	-	NA
MetaNumerics_NormalQuantile	179.68 ns	0.502 ns	0.028 ns	0.074	0.00	-	-	NA

Lodestar.Stats.Benchmarks.DistributionTailBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
ChiSquaredSfOneDf	31.05 ns	0.436 ns	0.024 ns	1.00	-	NA
ChiSquaredSfFourDf	13.85 ns	0.779 ns	0.043 ns	0.45	-	NA
ChiSquaredSfThreeDfFarTail	43.84 ns	0.128 ns	0.007 ns	1.41	-	NA
ChiSquaredSfHundredDf	76.87 ns	4.268 ns	0.234 ns	2.48	-	NA
ChiSquaredSfFractionalDf	118.43 ns	2.746 ns	0.151 ns	3.81	-	NA
NormalQuantile	28.75 ns	0.129 ns	0.007 ns	0.93	-	NA

Lodestar.Stats.Benchmarks.GlmBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Regressors	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Allocated	Alloc Ratio
Lodestar_Glm	200	1	35.02 μs	3.472 μs	0.190 μs	1.00	0.01	0.7324	-	12.13 KB	1.00
Accord_Glm	200	1	50.85 μs	14.241 μs	0.781 μs	1.45	0.02	7.2021	0.3052	117.67 KB	9.70
Lodestar_Glm	200	3	42.59 μs	2.056 μs	0.113 μs	1.00	0.00	1.1597	0.0610	19.38 KB	1.00
Accord_Glm	200	3	102.73 μs	5.060 μs	0.277 μs	2.41	0.01	10.2539	0.6104	168.44 KB	8.69
Lodestar_Glm	2000	1	370.96 μs	6.210 μs	0.340 μs	1.00	0.00	6.3477	0.4883	110.56 KB	1.00
Accord_Glm	2000	1	371.20 μs	52.102 μs	2.856 μs	1.00	0.01	53.7109	16.1133	881.77 KB	7.98
Lodestar_Glm	2000	3	411.10 μs	38.199 μs	2.094 μs	1.00	0.01	10.2539	-	174.06 KB	1.00
Accord_Glm	2000	3	739.40 μs	54.887 μs	3.009 μs	1.80	0.01	77.1484	33.2031	1265.51 KB	7.27

Lodestar.Stats.Benchmarks.GlmNegativeBinomialBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	1000	276.4 μs	20.94 μs	1.15 μs	3.4180	-	-	58.2 KB
Fit	10000	3,059.5 μs	105.63 μs	5.79 μs	97.6563	97.6563	97.6563	550.45 KB
Fit	100000	31,815.2 μs	2,651.37 μs	145.33 μs	437.5000	437.5000	437.5000	5474.06 KB

Lodestar.Stats.Benchmarks.GlmOffsetBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Poisson	200	30.21 μs	4.064 μs	0.223 μs	1.00	0.01	0.7324	-	-	12.13 KB	1.00
PoissonWithExposure	200	60.69 μs	12.708 μs	0.697 μs	2.01	0.02	1.2207	-	-	20.52 KB	1.69
Poisson	20000	3,998.64 μs	216.307 μs	11.857 μs	1.00	0.00	328.1250	328.1250	328.1250	1097.71 KB	1.00
PoissonWithExposure	20000	7,616.45 μs	789.580 μs	43.280 μs	1.90	0.01	546.8750	546.8750	546.8750	1877.96 KB	1.71

Lodestar.Stats.Benchmarks.GlmPoissonBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	MeanCount	Mean	Error	StdDev	Gen0	Gen1	Allocated
Fit	5	304.7 μs	157.11 μs	8.61 μs	6.3477	0.4883	110.56 KB
Fit	50000	306.0 μs	9.11 μs	0.50 μs	6.3477	0.4883	110.56 KB
Fit	5000000	307.2 μs	34.30 μs	1.88 μs	6.3477	0.4883	110.56 KB

Lodestar.Stats.Benchmarks.GlsBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Gls	50	50.69 μs	6.661 μs	0.365 μs	1.00	0.01	1.6479	0.0610	-	27.1 KB	1.00
MathNet_Gls	50	106.82 μs	15.684 μs	0.860 μs	2.11	0.02	1.9531	0.1221	-	33.02 KB	1.22
Lodestar_Gls	200	697.84 μs	52.052 μs	2.853 μs	1.00	0.01	99.6094	99.6094	99.6094	336.57 KB	1.00
MathNet_Gls	200	1,957.68 μs	584.190 μs	32.021 μs	2.81	0.04	97.6563	97.6563	97.6563	370.09 KB	1.10
Lodestar_Gls	500	6,864.98 μs	965.753 μs	52.936 μs	1.00	0.01	320.3125	320.3125	320.3125	2010.21 KB	1.00
MathNet_Gls	500	23,041.40 μs	8,892.739 μs	487.441 μs	3.36	0.07	281.2500	281.2500	281.2500	2227.2 KB	1.11
Lodestar_Gls	1000	51,756.89 μs	2,705.446 μs	148.295 μs	1.00	0.00	100.0000	100.0000	100.0000	7924.17 KB	1.00
MathNet_Gls	1000	141,294.34 μs	111,292.465 μs	6,100.316 μs	2.73	0.10	-	-	-	8846 KB	1.12

Lodestar.Stats.Benchmarks.HacClusterBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Lags	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Nonrobust	100	4	37.385 μs	0.9716 μs	0.0533 μs	1.00	0.00	0.1221	-	-	2.65 KB	1.00
Hc0	100	4	10.741 μs	6.7387 μs	0.3694 μs	0.29	0.01	0.5341	-	-	8.8 KB	3.32
Hac	100	4	20.034 μs	2.3301 μs	0.1277 μs	0.54	0.00	0.7629	-	-	12.95 KB	4.89
Cluster	100	4	9.159 μs	0.4580 μs	0.0251 μs	0.24	0.00	0.5951	-	-	9.76 KB	3.68
Nonrobust	10000	4	461.294 μs	38.3742 μs	2.1034 μs	1.00	0.01	4.3945	0.9766	-	79.99 KB	1.00
Hc0	10000	4	890.402 μs	118.1500 μs	6.4762 μs	1.93	0.01	124.0234	124.0234	124.0234	472.94 KB	5.91
Hac	10000	4	1,949.544 μs	102.7058 μs	5.6296 μs	4.23	0.02	246.0938	246.0938	246.0938	863.89 KB	10.80
Cluster	10000	4	767.062 μs	29.3895 μs	1.6109 μs	1.66	0.01	124.0234	124.0234	124.0234	565.22 KB	7.07

Lodestar.Stats.Benchmarks.KolmogorovDurbinBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Asymptotic	100	7.777 μs	0.5679 μs	0.0311 μs	0.2747	-	-	4.73 KB
Asymptotic	280	96.404 μs	19.0445 μs	1.0439 μs	1.4648	-	-	24.7 KB
Asymptotic	20000	1,301.736 μs	148.0936 μs	8.1175 μs	99.6094	99.6094	99.6094	352.77 KB

Lodestar.Stats.Benchmarks.KsAutoBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	Gen0	Allocated	Alloc Ratio
Before_Asymptotic	1000	31.27 μs	3.796 μs	0.208 μs	1.00	0.9155	15.73 KB	1.00
After_Auto	1000	35.32 μs	2.926 μs	0.160 μs	1.13	0.9155	15.67 KB	1.00
Before_Asymptotic	10000	1,985.53 μs	82.889 μs	4.543 μs	1.00	7.8125	156.3 KB	1.00
After_Auto	10000	1,461.53 μs	72.159 μs	3.955 μs	0.74	7.8125	156.3 KB	1.00

Lodestar.Stats.Benchmarks.KsExactTableBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Gen0	Allocated
TwoSidedExact	999x1001	3,431.91 μs	381.164 μs	20.893 μs	-	39.28 KB
TwoSidedExact	99x101	46.92 μs	3.740 μs	0.205 μs	0.2441	4.12 KB

Lodestar.Stats.Benchmarks.LeastSquaresRoutingBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Regressors	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Estimate	4	142.1 μs	32.27 μs	1.77 μs	1.00	0.02	49.8047	49.8047	49.8047	188.26 KB	1.00
Lodestar_Fit	4	203.8 μs	4.10 μs	0.22 μs	1.43	0.02	1.9531	-	-	33.12 KB	0.18
MathNet_Qr	4	660.1 μs	258.03 μs	14.14 μs	4.65	0.10	190.4297	190.4297	190.4297	706.08 KB	3.75
Lodestar_Estimate	250	62,179.2 μs	46,550.84 μs	2,551.61 μs	1.00	0.05	375.0000	375.0000	375.0000	8865.83 KB	1.00
Lodestar_Fit	250	297,560.1 μs	14,383.71 μs	788.42 μs	4.79	0.17	-	-	-	3499.3 KB	0.39
MathNet_Qr	250	490,262.8 μs	261,082.00 μs	14,310.79 μs	7.89	0.34	-	-	-	32662.32 KB	3.68

Lodestar.Stats.Benchmarks.MannWhitneyExactBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Exact	141x141	545.0 ms	50.44 ms	2.77 ms	1000.0000	1000.0000	1000.0000	43.23 MB
Exact	8x2500	441.5 ms	22.34 ms	1.22 ms	-	-	-	2.9 MB

Lodestar.Stats.Benchmarks.MathNetInteropBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Size	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Diagonal	1000	4.264 μs	2.099 μs	0.1151 μs	0.9613	0.0381	-	15.75 KB
SparseDense	1000	1,573.114 μs	59.579 μs	3.2657 μs	5.8594	-	-	126.59 KB
ToSparseSorted	1000	23.155 μs	2.350 μs	0.1288 μs	7.4768	0.9155	-	122.69 KB
ToSparseReversed	1000	92.845 μs	2.759 μs	0.1513 μs	7.6904	2.4414	-	126.63 KB
Diagonal	4000	15.822 μs	1.840 μs	0.1009 μs	3.8147	0.2136	-	62.63 KB
SparseDense	4000	26,194.222 μs	276.324 μs	15.1463 μs	-	-	-	1909.55 KB
ToSparseSorted	4000	528.651 μs	476.153 μs	26.0996 μs	7.8125	7.8125	7.8125	1893.94 KB
ToSparseReversed	4000	1,343.936 μs	89.162 μs	4.8873 μs	7.8125	7.8125	7.8125	1909.57 KB

Lodestar.Stats.Benchmarks.MetaNumericsStatsBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar_StudentT	100	4,601.9 ns	1,060.57 ns	58.13 ns	1.00	0.02	-	64 B	1.00
MetaNumerics_StudentT	100	2,200.9 ns	13.54 ns	0.74 ns	0.48	0.01	0.0038	104 B	1.62
Lodestar_MannWhitney	100	2,330.1 ns	504.44 ns	27.65 ns	0.51	0.01	-	-	0.00
MetaNumerics_MannWhitney	100	3,891.5 ns	753.96 ns	41.33 ns	0.85	0.01	0.0687	1176 B	18.38
Lodestar_KruskalWallis	100	7,715.9 ns	1,967.47 ns	107.84 ns	1.68	0.03	-	32 B	0.50
MetaNumerics_KruskalWallis	100	9,681.3 ns	1,430.04 ns	78.39 ns	2.10	0.03	0.1068	1896 B	29.62
Lodestar_KolmogorovSmirnov	100	2,625.7 ns	55.04 ns	3.02 ns	0.57	0.01	0.0954	1648 B	25.75
MetaNumerics_KolmogorovSmirnov	100	4,061.8 ns	754.07 ns	41.33 ns	0.88	0.01	0.0687	1168 B	18.25
Lodestar_OneWayAnova	100	1,665.2 ns	95.83 ns	5.25 ns	0.36	0.00	0.0019	32 B	0.50
MetaNumerics_OneWayAnova	100	3,159.3 ns	90.34 ns	4.95 ns	0.69	0.01	0.0420	744 B	11.62
Lodestar_Wilcoxon	100	1,141.6 ns	285.59 ns	15.65 ns	0.25	0.00	0.0477	824 B	12.88
MetaNumerics_Wilcoxon	100	1,076.8 ns	263.66 ns	14.45 ns	0.23	0.00	0.0572	976 B	15.25
Lodestar_FisherExact	100	459.6 ns	6.74 ns	0.37 ns	0.10	0.00	-	-	0.00
MetaNumerics_FisherExact	100	1,543.4 ns	46.58 ns	2.55 ns	0.34	0.00	0.0553	944 B	14.75
Lodestar_ChiSquare	100	117.8 ns	5.46 ns	0.30 ns	0.03	0.00	0.0100	168 B	2.62
MetaNumerics_ChiSquare	100	697.0 ns	47.78 ns	2.62 ns	0.15	0.00	0.0572	968 B	15.12
Lodestar_StudentT	10000	59,955.8 ns	5,767.49 ns	316.14 ns	1.000	0.01	-	64 B	1.00
MetaNumerics_StudentT	10000	178,730.7 ns	1,813.27 ns	99.39 ns	2.981	0.01	-	104 B	1.62
Lodestar_MannWhitney	10000	686,331.3 ns	25,017.35 ns	1,371.29 ns	11.447	0.06	-	-	0.00
MetaNumerics_MannWhitney	10000	2,100,516.6 ns	105,629.08 ns	5,789.89 ns	35.035	0.18	3.9063	80382 B	1,255.97
Lodestar_KruskalWallis	10000	1,469,606.2 ns	99,978.30 ns	5,480.15 ns	24.512	0.14	-	35 B	0.55
MetaNumerics_KruskalWallis	10000	3,583,938.3 ns	196,158.61 ns	10,752.12 ns	59.777	0.31	3.9063	120702 B	1,885.97
Lodestar_KolmogorovSmirnov	10000	1,460,150.7 ns	98,543.76 ns	5,401.52 ns	24.354	0.14	7.8125	160051 B	2,500.80
MetaNumerics_KolmogorovSmirnov	10000	2,161,099.2 ns	89,185.33 ns	4,888.55 ns	36.046	0.18	3.9063	80374 B	1,255.84
Lodestar_OneWayAnova	10000	85,483.3 ns	1,345.16 ns	73.73 ns	1.426	0.01	-	32 B	0.50
MetaNumerics_OneWayAnova	10000	268,490.3 ns	5,934.88 ns	325.31 ns	4.478	0.02	-	745 B	11.64
Lodestar_Wilcoxon	10000	190,207.8 ns	19,346.64 ns	1,060.45 ns	3.173	0.02	4.6387	80056 B	1,250.88
MetaNumerics_Wilcoxon	10000	386,668.4 ns	145,417.38 ns	7,970.82 ns	6.449	0.12	4.3945	80177 B	1,252.77
Lodestar_FisherExact	10000	460.6 ns	20.69 ns	1.13 ns	0.008	0.00	-	-	0.00
MetaNumerics_FisherExact	10000	1,564.6 ns	473.37 ns	25.95 ns	0.026	0.00	0.0553	944 B	14.75
Lodestar_ChiSquare	10000	118.9 ns	8.58 ns	0.47 ns	0.002	0.00	0.0100	168 B	2.62
MetaNumerics_ChiSquare	10000	697.5 ns	70.12 ns	3.84 ns	0.012	0.00	0.0572	968 B	15.12

Lodestar.Stats.Benchmarks.MultinomialLogitBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Allocated	Alloc Ratio
Lodestar	200	135.6 μs	5.92 μs	0.32 μs	1.00	0.00	0.9766	-	19.24 KB	1.00
Accord	200	892.2 μs	13.68 μs	0.75 μs	6.58	0.01	88.8672	7.8125	1466.83 KB	76.23
Lodestar	2000	1,312.2 μs	193.17 μs	10.59 μs	1.00	0.01	7.8125	-	131.74 KB	1.00
Accord	2000	7,104.2 μs	1,293.60 μs	70.91 μs	5.41	0.06	734.3750	359.3750	12015.21 KB	91.20

Lodestar.Stats.Benchmarks.OlsBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Ols	100	36.59 μs	0.462 μs	0.025 μs	1.00	0.00	0.1221	-	-	2.65 KB	1.00
Accord_Ols	100	265.50 μs	20.770 μs	1.138 μs	7.26	0.03	2.4414	-	-	43 KB	16.24
MathNet_Ols	100	30.88 μs	1.954 μs	0.107 μs	0.84	0.00	2.1973	0.0610	-	35.66 KB	13.46
Lodestar_Ols	10000	458.86 μs	17.757 μs	0.973 μs	1.00	0.00	4.3945	0.9766	-	79.99 KB	1.00
Accord_Ols	10000	4,561.71 μs	946.986 μs	51.908 μs	9.94	0.10	210.9375	187.5000	-	3523.47 KB	44.05
MathNet_Ols	10000	1,688.32 μs	207.871 μs	11.394 μs	3.68	0.02	441.4063	441.4063	441.4063	1765.79 KB	22.07

Lodestar.Stats.Benchmarks.QuantileBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Median	Ratio	RatioSD	Allocated	Alloc Ratio
NormalQuantile	0.0550 ns	0.4986 ns	0.0273 ns	0.0394 ns	1.15	0.64	-	NA
NormalQuantileFarTail	0.0562 ns	0.3409 ns	0.0187 ns	0.0493 ns	1.17	0.53	-	NA
StudentQuantile	14,224.2219 ns	2,304.6747 ns	126.3270 ns	14,292.7388 ns	296,352.22	99,099.41	-	NA
StudentQuantileCauchyFarTail	2,329.8424 ns	23.2012 ns	1.2717 ns	2,329.1242 ns	48,540.72	16,227.19	-	NA

Lodestar.Stats.Benchmarks.RankTestBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Ties	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
KruskalWallisTest	10000	False	1,486.7 μs	38.08 μs	2.09 μs	-	-	-	82 B
WilcoxonPaired	10000	False	419.2 μs	49.99 μs	2.74 μs	4.3945	-	-	80057 B
KruskalWallisManyGroups	10000	False	618.0 μs	303.10 μs	16.61 μs	83.0078	83.0078	83.0078	741656 B
WilcoxonPooledDifferences	10000	False	1,263.0 μs	31.32 μs	1.72 μs	-	-	-	-
MannWhitneyTest	10000	False	691.9 μs	33.26 μs	1.82 μs	-	-	-	-
KruskalWallisTest	10000	True	620.8 μs	12.52 μs	0.69 μs	-	-	-	81 B
WilcoxonPaired	10000	True	244.8 μs	29.28 μs	1.61 μs	4.6387	-	-	80056 B
KruskalWallisManyGroups	10000	True	517.7 μs	14.12 μs	0.77 μs	83.0078	83.0078	83.0078	741656 B
WilcoxonPooledDifferences	10000	True	781.0 μs	12.50 μs	0.69 μs	-	-	-	-
MannWhitneyTest	10000	True	398.4 μs	14.52 μs	0.80 μs	-	-	-	-
KruskalWallisTest	100000	False	14,915.0 μs	1,175.43 μs	64.43 μs	-	-	-	-
WilcoxonPaired	100000	False	4,484.1 μs	99.16 μs	5.44 μs	101.5625	101.5625	101.5625	800164 B
KruskalWallisManyGroups	100000	False	4,528.7 μs	1,515.33 μs	83.06 μs	507.8125	507.8125	507.8125	5062691 B
WilcoxonPooledDifferences	100000	False	17,198.1 μs	1,290.14 μs	70.72 μs	468.7500	468.7500	468.7500	19462711 B
MannWhitneyTest	100000	False	7,193.4 μs	763.15 μs	41.83 μs	-	-	-	-
KruskalWallisTest	100000	True	6,072.4 μs	408.86 μs	22.41 μs	-	-	-	88 B
WilcoxonPaired	100000	True	2,798.7 μs	115.21 μs	6.32 μs	113.2813	113.2813	113.2813	800172 B
KruskalWallisManyGroups	100000	True	3,335.4 μs	338.28 μs	18.54 μs	523.4375	523.4375	523.4375	5062699 B
WilcoxonPooledDifferences	100000	True	13,623.3 μs	3,866.66 μs	211.94 μs	531.2500	531.2500	531.2500	19281691 B
MannWhitneyTest	100000	True	3,839.6 μs	158.47 μs	8.69 μs	-	-	-	-

Lodestar.Stats.Benchmarks.RobustCovarianceBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Covariance	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	100	Nonrobust	36.68 μs	2.279 μs	0.125 μs	0.1221	-	-	2.6 KB
Fit	100	Hc0	10.78 μs	1.457 μs	0.080 μs	0.5341	-	-	8.75 KB
Fit	100	Hc1	10.89 μs	0.873 μs	0.048 μs	0.5341	-	-	8.75 KB
Fit	100	Hc2	12.05 μs	0.071 μs	0.004 μs	0.5798	-	-	9.62 KB
Fit	100	Hc3	11.93 μs	0.240 μs	0.013 μs	0.5798	-	-	9.62 KB
Fit	10000	Nonrobust	452.90 μs	19.627 μs	1.076 μs	4.3945	0.9766	-	79.95 KB
Fit	10000	Hc0	895.39 μs	30.469 μs	1.670 μs	124.0234	124.0234	124.0234	472.89 KB
Fit	10000	Hc1	892.15 μs	57.603 μs	3.157 μs	124.0234	124.0234	124.0234	472.89 KB
Fit	10000	Hc2	1,079.66 μs	828.688 μs	45.423 μs	123.0469	123.0469	123.0469	551.11 KB
Fit	10000	Hc3	1,059.66 μs	213.410 μs	11.698 μs	123.0469	123.0469	123.0469	551.11 KB

Lodestar.Stats.Benchmarks.SerialCorrelationBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
LodestarAutocorrelation	200	4.395 μs	0.3710 μs	0.0203 μs	1.00	0.01	0.1526	2624 B	1.00
CortexAutocorrelation	200	11.539 μs	0.4925 μs	0.0270 μs	2.63	0.01	-	192 B	0.07
LodestarPartialAutocorrelation	200	5.161 μs	0.2404 μs	0.0132 μs	1.17	0.01	0.1678	2816 B	1.07
CortexPartialAutocorrelation	200	12.165 μs	0.4394 μs	0.0241 μs	2.77	0.01	0.0458	768 B	0.29
LodestarLjungBox	200	4.971 μs	0.4211 μs	0.0231 μs	1.13	0.01	0.1373	2344 B	0.89
CortexLjungBox	200	11.665 μs	0.4122 μs	0.0226 μs	2.65	0.01	-	224 B	0.09
LodestarAutocorrelation	2000	48.781 μs	1.4305 μs	0.0784 μs	1.00	0.00	0.9766	17024 B	1.00
CortexAutocorrelation	2000	120.995 μs	1.7669 μs	0.0968 μs	2.48	0.00	-	192 B	0.01
LodestarPartialAutocorrelation	2000	49.592 μs	0.6844 μs	0.0375 μs	1.02	0.00	0.9766	17216 B	1.01
CortexPartialAutocorrelation	2000	121.715 μs	4.2076 μs	0.2306 μs	2.50	0.01	-	768 B	0.05
LodestarLjungBox	2000	49.241 μs	0.6126 μs	0.0336 μs	1.01	0.00	0.9766	16744 B	0.98
CortexLjungBox	2000	122.325 μs	41.2171 μs	2.2592 μs	2.51	0.04	-	224 B	0.01

Lodestar.Stats.Benchmarks.StationarityBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
LodestarAugmentedDickeyFuller	200	7.520 μs	1.3971 μs	0.0766 μs	1.00	0.01	1.4648	0.0458	-	23.99 KB	1.00
LodestarAdfAutolag	200	30.301 μs	1.6230 μs	0.0890 μs	4.03	0.04	4.0588	0.2441	-	66.56 KB	2.77
CortexAugmentedDickeyFuller	200	26.730 μs	7.8007 μs	0.4276 μs	3.55	0.06	0.9766	0.0305	-	16.18 KB	0.67
LodestarKpss	200	3.876 μs	0.1763 μs	0.0097 μs	0.52	0.00	0.0992	-	-	1.69 KB	0.07
CortexKpss	200	4.056 μs	0.7058 μs	0.0387 μs	0.54	0.01	0.2060	-	-	3.43 KB	0.14
LodestarDecompose	200	1.538 μs	0.4980 μs	0.0273 μs	0.20	0.00	0.3967	0.0038	-	6.5 KB	0.27
CortexDecompose	200	2.936 μs	0.8623 μs	0.0473 μs	0.39	0.01	0.4997	0.0076	-	8.16 KB	0.34
LodestarAugmentedDickeyFuller	2000	117.502 μs	36.1880 μs	1.9836 μs	1.00	0.02	30.2734	30.2734	30.2734	234.95 KB	1.00
LodestarAdfAutolag	2000	999.996 μs	92.9118 μs	5.0928 μs	8.51	0.13	248.0469	248.0469	248.0469	1001.27 KB	4.26
CortexAugmentedDickeyFuller	2000	295.414 μs	30.9536 μs	1.6967 μs	2.51	0.04	30.2734	30.2734	30.2734	142.76 KB	0.61
LodestarKpss	2000	95.805 μs	2.6006 μs	0.1425 μs	0.82	0.01	0.8545	-	-	15.75 KB	0.07
CortexKpss	2000	97.584 μs	3.2444 μs	0.1778 μs	0.83	0.01	1.8311	-	-	31.55 KB	0.13
LodestarDecompose	2000	14.404 μs	2.5213 μs	0.1382 μs	0.12	0.00	3.8300	0.4120	-	62.75 KB	0.27
CortexDecompose	2000	28.893 μs	1.9534 μs	0.1071 μs	0.25	0.00	4.7913	0.4578	-	78.48 KB	0.33

Lodestar.Stats.Benchmarks.StatsBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
LodestarWelchT	100	6,116.0 ns	737.12 ns	40.40 ns	1.00	0.01	-	-	-	64 B	1.00
AccordWelchT	100	38,431.4 ns	574.22 ns	31.47 ns	6.28	0.04	-	-	-	392 B	6.12
LodestarMannWhitney	100	2,160.6 ns	181.86 ns	9.97 ns	0.35	0.00	-	-	-	-	0.00
AccordMannWhitney	100	23,775.4 ns	863.24 ns	47.32 ns	3.89	0.02	1.3733	0.0305	-	23336 B	364.62
LodestarChiSquare	100	118.8 ns	33.48 ns	1.83 ns	0.02	0.00	0.0100	-	-	168 B	2.62
AccordChiSquare	100	221.8 ns	12.53 ns	0.69 ns	0.04	0.00	0.0100	-	-	168 B	2.62
LodestarWelchT	10000	53,526.8 ns	315.12 ns	17.27 ns	1.000	0.00	-	-	-	64 B	1.00
AccordWelchT	10000	141,119.0 ns	9,439.72 ns	517.42 ns	2.636	0.01	-	-	-	392 B	6.12
LodestarMannWhitney	10000	694,243.7 ns	15,747.61 ns	863.18 ns	12.970	0.01	-	-	-	-	0.00
AccordMannWhitney	10000	12,696,886.8 ns	2,728,287.46 ns	149,546.64 ns	237.206	2.42	234.3750	234.3750	234.3750	2241217 B	35,019.02
LodestarChiSquare	10000	118.3 ns	12.77 ns	0.70 ns	0.002	0.00	0.0100	-	-	168 B	2.62
AccordChiSquare	10000	215.3 ns	11.95 ns	0.65 ns	0.004	0.00	0.0100	-	-	168 B	2.62

Lodestar.Stats.Benchmarks.VarianceAndFitBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	GroupSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Allocated	Alloc Ratio
Lodestar_Levene	100	2,857.0 ns	293.95 ns	16.11 ns	1.00	0.01	0.3014	-	5072 B	1.000
Accord_Levene	100	4,100.2 ns	432.20 ns	23.69 ns	1.44	0.01	0.3204	-	5480 B	1.080
Lodestar_Bartlett	100	604.1 ns	21.28 ns	1.17 ns	0.21	0.00	0.0048	-	80 B	0.016
Accord_Bartlett	100	1,185.1 ns	64.91 ns	3.56 ns	0.41	0.00	0.0095	-	168 B	0.033
Lodestar_Binomial	100	810.3 ns	19.46 ns	1.07 ns	0.28	0.00	0.0029	-	48 B	0.009
Accord_Binomial	100	4,565.2 ns	141.80 ns	7.77 ns	1.60	0.01	0.1297	-	2216 B	0.437
Lodestar_AndersonDarling	100	9,069.0 ns	1,696.98 ns	93.02 ns	3.17	0.03	0.0458	-	1000 B	0.197
Accord_AndersonDarling	100	6,238.4 ns	165.64 ns	9.08 ns	2.18	0.01	0.0610	-	1096 B	0.216
Lodestar_ClopperPearson	100	4,311.9 ns	122.99 ns	6.74 ns	1.51	0.01	-	-	48 B	0.009
Lodestar_Levene	10000	389,120.1 ns	26,693.13 ns	1,463.14 ns	1.000	0.00	28.3203	-	480272 B	1.000
Accord_Levene	10000	1,347,393.2 ns	62,665.94 ns	3,434.93 ns	3.463	0.01	27.3438	5.8594	480681 B	1.001
Lodestar_Bartlett	10000	56,296.6 ns	1,329.44 ns	72.87 ns	0.145	0.00	-	-	80 B	0.000
Accord_Bartlett	10000	112,502.2 ns	6,562.99 ns	359.74 ns	0.289	0.00	-	-	168 B	0.000
Lodestar_Binomial	10000	1,032.5 ns	65.13 ns	3.57 ns	0.003	0.00	0.0019	-	48 B	0.000
Accord_Binomial	10000	887,048.7 ns	93,462.09 ns	5,122.97 ns	2.280	0.01	11.7188	2.9297	200217 B	0.417
Lodestar_AndersonDarling	10000	1,441,847.4 ns	118,221.95 ns	6,480.14 ns	3.705	0.02	3.9063	-	80201 B	0.167
Accord_AndersonDarling	10000	NA	NA	NA	?	?	NA	NA	NA	?
Lodestar_ClopperPearson	10000	11,046.1 ns	2,203.60 ns	120.79 ns	0.028	0.00	-	-	48 B	0.000

Benchmarks with issues: VarianceAndFitBenchmarks.Accord_AndersonDarling: ShortRun(IterationCount=3, LaunchCount=1, WarmupCount=3) [GroupSize=10000]

Lodestar.Stats.Benchmarks.VectorAutoregressionBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Variables	Lags	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	500	2	1	14.99 μs	2.408 μs	0.132 μs	2.9602	0.3510	-	48.37 KB
Fit	500	2	4	35.81 μs	2.199 μs	0.121 μs	5.8594	1.4038	-	96.13 KB
Fit	500	5	1	51.27 μs	6.390 μs	0.350 μs	6.6528	1.6479	-	109.52 KB
Fit	500	5	4	158.64 μs	37.033 μs	2.030 μs	14.1602	6.8359	-	233.92 KB
Fit	5000	2	1	215.03 μs	11.573 μs	0.634 μs	73.9746	73.9746	73.9746	470.29 KB
Fit	5000	2	4	555.95 μs	32.174 μs	1.764 μs	221.6797	221.6797	221.6797	940.03 KB
Fit	5000	5	1	742.05 μs	25.079 μs	1.375 μs	142.5781	142.5781	142.5781	1058.83 KB
Fit	5000	5	4	2,319.70 μs	73.536 μs	4.031 μs	496.0938	496.0938	496.0938	2238.16 KB

Lodestar.Stats.Benchmarks.WeightedLeastSquaresBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Wls	200	45.09 μs	1.718 μs	0.094 μs	1.00	0.00	0.1831	-	-	3.43 KB	1.00
MathNet_Wls	200	18.90 μs	0.333 μs	0.018 μs	0.42	0.00	3.6926	0.1831	-	58.29 KB	16.99
Lodestar_Wls	2000	136.16 μs	10.785 μs	0.591 μs	1.00	0.01	0.9766	-	-	17.49 KB	1.00
MathNet_Wls	2000	121.05 μs	21.304 μs	1.168 μs	0.89	0.01	33.2031	13.6719	-	508.29 KB	29.06
Lodestar_Wls	20000	931.50 μs	14.989 μs	0.822 μs	1.00	0.00	46.8750	46.8750	46.8750	158.51 KB	1.00
MathNet_Wls	20000	1,634.31 μs	2,243.496 μs	122.974 μs	1.75	0.11	1019.5313	1019.5313	1019.5313	5027.72 KB	31.72
Lodestar_Wls	200000	9,136.69 μs	312.581 μs	17.134 μs	1.00	0.00	93.7500	93.7500	93.7500	1564.71 KB	1.00
MathNet_Wls	200000	18,766.16 μs	31,228.612 μs	1,711.746 μs	2.05	0.16	718.7500	718.7500	718.7500	50025.1 KB	31.97

Lodestar.Survival.Benchmarks.CoxBenchmarks-report-github

As of 2026-11-12, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 11.0.100
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	Covariates	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	1000	2	331.1 μs	20.53 μs	1.13 μs	8.3008	0.9766	-	140.32 KB
Fit	1000	8	2,041.3 μs	205.33 μs	11.25 μs	11.7188	-	-	201.32 KB
Fit	10000	2	5,567.9 μs	332.36 μs	18.22 μs	164.0625	164.0625	164.0625	1248.62 KB
Fit	10000	8	18,695.6 μs	6,907.10 μs	378.60 μs	281.2500	250.0000	250.0000	1737.31 KB

Lodestar.Survival.Benchmarks.SurvivalBenchmarks-report-github

As of 2026-11-12, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 11.0.100
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleSize	DistinctPercent	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
KaplanMeierEstimate	1000	4	16.27 μs	2.767 μs	0.152 μs	1.00	0.01	0.9766	-	-	15.98 KB	1.00
NelsonAalenEstimate	1000	4	15.02 μs	0.702 μs	0.038 μs	0.92	0.01	0.9308	0.0153	-	15.27 KB	0.96
LogRankTest	1000	4	21.66 μs	3.509 μs	0.192 μs	1.33	0.01	1.6174	0.0610	-	26.84 KB	1.68
KaplanMeierEstimate	1000	100	59.88 μs	8.940 μs	0.490 μs	1.00	0.01	4.3945	0.4883	-	72 KB	1.00
NelsonAalenEstimate	1000	100	22.88 μs	1.454 μs	0.080 μs	0.38	0.00	3.7842	0.3967	-	62.2 KB	0.86
LogRankTest	1000	100	45.50 μs	0.990 μs	0.054 μs	0.76	0.01	2.9907	0.3052	-	49.57 KB	0.69
KaplanMeierEstimate	100000	4	9,941.86 μs	160.039 μs	8.772 μs	1.00	0.00	328.1250	312.5000	312.5000	1517.93 KB	1.00
NelsonAalenEstimate	100000	4	9,647.80 μs	1,281.320 μs	70.233 μs	0.97	0.01	328.1250	312.5000	312.5000	1455.34 KB	0.96
LogRankTest	100000	4	11,016.75 μs	694.078 μs	38.045 μs	1.11	0.00	484.3750	484.3750	484.3750	2598.82 KB	1.71
KaplanMeierEstimate	100000	100	16,903.37 μs	297.976 μs	16.333 μs	1.00	0.00	1125.0000	1093.7500	968.7500	6650.01 KB	1.00
NelsonAalenEstimate	100000	100	13,593.09 μs	3,595.161 μs	197.063 μs	0.80	0.01	875.0000	828.1250	718.7500	5660.24 KB	0.85
LogRankTest	100000	100	17,858.19 μs	6,967.616 μs	381.919 μs	1.06	0.02	500.0000	500.0000	500.0000	4916.39 KB	0.74

Lodestar.Text.Benchmarks.AddedTokenScanBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Allocated
ChatTemplate	97.60 ms	0.956 ms	0.052 ms	2000.0000	33.3 MB
ProseWithoutAddedTokens	82.51 ms	6.132 ms	0.336 ms	1714.2857	28.47 MB

Lodestar.Text.Benchmarks.AgglomerativeIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Fit	500	average	1,953.9 μs	140.98 μs	7.73 μs	1.00	0.00	246.0938	246.0938	246.0938	1022.55 KB	1.00
Aglomera_GetClustering	500	average	89,115.5 μs	5,192.60 μs	284.62 μs	45.61	0.20	3666.6667	1500.0000	833.3333	59880.37 KB	58.56
Lodestar_Fit	500	complete	1,880.0 μs	406.67 μs	22.29 μs	1.00	0.01	248.0469	248.0469	248.0469	1022.4 KB	1.00
Aglomera_GetClustering	500	complete	99,685.4 μs	48,813.89 μs	2,675.65 μs	53.03	1.35	5000.0000	1800.0000	800.0000	83111.74 KB	81.29
Lodestar_Fit	500	single	776.7 μs	81.88 μs	4.49 μs	1.00	0.01	0.9766	-	-	30.34 KB	1.00
Aglomera_GetClustering	500	single	143,517.6 μs	29,425.66 μs	1,612.92 μs	184.78	2.02	4750.0000	1750.0000	750.0000	81289.96 KB	2,678.88
Lodestar_Fit	500	ward	2,232.9 μs	555.77 μs	30.46 μs	1.00	0.02	246.0938	246.0938	246.0938	1022.4 KB	1.00
Aglomera_GetClustering	500	ward	80,153.9 μs	7,401.13 μs	405.68 μs	35.90	0.45	3875.0000	1750.0000	875.0000	64191.22 KB	62.78
Lodestar_Fit	1500	average	15,593.0 μs	2,632.51 μs	144.30 μs	1.00	0.01	500.0000	500.0000	500.0000	8925.16 KB	1.00
Aglomera_GetClustering	1500	average	1,792,095.7 μs	253,935.42 μs	13,919.06 μs	114.94	1.20	35000.0000	10000.0000	4000.0000	537152.06 KB	60.18
Lodestar_Fit	1500	complete	14,413.2 μs	5,807.93 μs	318.35 μs	1.00	0.03	500.0000	500.0000	500.0000	8925 KB	1.00
Aglomera_GetClustering	1500	complete	1,884,224.1 μs	133,581.81 μs	7,322.07 μs	130.77	2.53	48000.0000	12000.0000	4000.0000	747352.87 KB	83.74
Lodestar_Fit	1500	single	6,376.1 μs	222.15 μs	12.18 μs	1.00	0.00	-	-	-	89.92 KB	1.00
Aglomera_GetClustering	1500	single	3,138,453.6 μs	160,454.76 μs	8,795.07 μs	492.22	1.45	47000.0000	12000.0000	4000.0000	732673.02 KB	8,147.88
Lodestar_Fit	1500	ward	18,186.5 μs	2,190.64 μs	120.08 μs	1.00	0.01	500.0000	500.0000	500.0000	8925.01 KB	1.00
Aglomera_GetClustering	1500	ward	1,613,604.3 μs	88,773.14 μs	4,865.96 μs	88.73	0.56	38000.0000	11000.0000	4000.0000	573621.47 KB	64.27

Lodestar.Text.Benchmarks.ArtifactLoadOnlyBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
CountLoadAsync2M	26.17 ms	22.549 ms	1.236 ms	750.0000	687.5000	343.7500	21.53 MB
CountLoadAsync4M	53.85 ms	110.222 ms	6.042 ms	1272.7273	1181.8182	454.5455	43.93 MB
CountLoad4M	45.92 ms	4.753 ms	0.261 ms	1363.6364	1272.7273	545.4545	43.92 MB

Lodestar.Text.Benchmarks.ArtifactTimeBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
CountFitLongDocuments	178,954.81 μs	28,965.88 μs	1,587.718 μs	-	-	-	54334.63 KB
CountSave240K	404.93 μs	24.56 μs	1.346 μs	9.2773	2.9297	2.9297	504.66 KB
CountSaveAsync240K	445.07 μs	41.25 μs	2.261 μs	11.2305	4.8828	4.8828	739.44 KB
TfidfSave45K	51.87 μs	25.62 μs	1.404 μs	8.3618	1.6479	-	137.13 KB
TfidfSaveAsync45K	52.42 μs	13.80 μs	0.756 μs	11.1084	-	-	181.88 KB
IndexSave942K	288.38 μs	52.54 μs	2.880 μs	7.3242	0.9766	0.9766	243.5 KB
IndexSaveAsync942K	289.21 μs	54.99 μs	3.014 μs	7.8125	0.9766	0.9766	243.62 KB
CountLoadAsync2M	27,629.03 μs	27,480.93 μs	1,506.323 μs	593.7500	562.5000	187.5000	22048.96 KB
CountLoadAsync4M	58,927.28 μs	42,409.03 μs	2,324.582 μs	1100.0000	1000.0000	300.0000	44983.53 KB
CountLoad4M	53,792.81 μs	14,457.11 μs	792.443 μs	1181.8182	1090.9091	363.6364	44976.8 KB

Lodestar.Text.Benchmarks.AsyncArtifactBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
TfidfSaveAsyncSmall	17.095 μs	41.0941 μs	2.2525 μs	1.6785	0.0916	-	27.43 KB
CountSaveAsyncWide	513.955 μs	141.8196 μs	7.7736 μs	35.1563	29.2969	29.2969	740.13 KB
HashingLoadUndeclared	5.520 μs	0.8645 μs	0.0474 μs	5.0964	-	-	83.87 KB
TfidfLoadUndeclared	28.194 μs	1.0449 μs	0.0573 μs	7.4158	0.8240	-	121.83 KB
CountLoadUndeclared24K	128.090 μs	7.7814 μs	0.4265 μs	16.1133	3.1738	-	264.67 KB
CountLoadUndeclared78K	882.854 μs	309.9054 μs	16.9870 μs	38.0859	37.1094	8.7891	656.32 KB
CountLoadUndeclared1M	11,200.079 μs	4,864.8579 μs	266.6593 μs	515.6250	500.0000	296.8750	10793.89 KB
CountSaveAsync1M	2,183.803 μs	2,457.8766 μs	134.7245 μs	109.3750	101.5625	101.5625	3162.11 KB
IndexSaveAsyncLongHead	3,438.531 μs	96.0010 μs	5.2621 μs	93.7500	89.8438	89.8438	2124.2 KB
CountLoadUndeclared2M	19,259.420 μs	1,797.5350 μs	98.5290 μs	1000.0000	968.7500	593.7500	22045.13 KB

Lodestar.Text.Benchmarks.BatchEmbeddingBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	CorpusSize	Mean	Error	StdDev	Ratio	Gen0	Gen1	Allocated	Alloc Ratio
UnitLoop	1	5.412 μs	0.4920 μs	0.0270 μs	1.00	0.1373	-	2.27 KB	1.00
EmbedBatch	1	5.891 μs	0.0995 μs	0.0055 μs	1.09	0.1602	-	2.7 KB	1.19
EmbedBatchBucketed	1	5.871 μs	0.0895 μs	0.0049 μs	1.08	0.1602	-	2.7 KB	1.19
UnitLoop	8	57.396 μs	7.4596 μs	0.4089 μs	1.00	1.7700	-	29.48 KB	1.00
EmbedBatch	8	21.769 μs	3.8469 μs	0.2109 μs	0.38	1.3428	-	22.27 KB	0.76
EmbedBatchBucketed	8	22.002 μs	1.8600 μs	0.1020 μs	0.38	1.3428	-	22.27 KB	0.76
UnitLoop	32	216.397 μs	18.6862 μs	1.0243 μs	1.00	6.5918	-	110.99 KB	1.00
EmbedBatch	32	80.187 μs	0.7624 μs	0.0418 μs	0.37	5.0049	0.2441	82.64 KB	0.74
EmbedBatchBucketed	32	70.383 μs	12.0873 μs	0.6625 μs	0.33	4.1504	0.1221	69.13 KB	0.62
UnitLoop	128	848.872 μs	185.5053 μs	10.1682 μs	1.00	26.3672	-	443.9 KB	1.00
EmbedBatch	128	320.062 μs	39.7549 μs	2.1791 μs	0.38	20.0195	2.4414	329.67 KB	0.74
EmbedBatchBucketed	128	273.408 μs	38.6394 μs	2.1180 μs	0.32	15.6250	1.9531	262.16 KB	0.59

Lodestar.Text.Benchmarks.BertNormalizerBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Text	Lowercase	Mean	Error	StdDev	Gen0	Allocated
Encode	Ascii	False	35.87 ms	3.802 ms	0.208 ms	714.2857	11.7 MB
Encode	Ascii	True	40.12 ms	6.564 ms	0.360 ms	692.3077	11.7 MB
Encode	Accented	False	34.48 ms	0.668 ms	0.037 ms	666.6667	11.69 MB
Encode	Accented	True	49.08 ms	3.455 ms	0.189 ms	1272.7273	20.38 MB
Encode	Cjk	False	40.46 ms	10.567 ms	0.579 ms	1307.6923	21.55 MB
Encode	Cjk	True	53.33 ms	2.604 ms	0.143 ms	1500.0000	24.48 MB

Lodestar.Text.Benchmarks.BkTreeBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Radius	Shape	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
LengthFilteredScan	1	clustered	132.88 ms	6.529 ms	0.358 ms	1.00	0.00	27.25 KB	1.00
TreeWithinDistance	1	clustered	65.79 ms	2.346 ms	0.129 ms	0.50	0.00	103.65 KB	3.80
LengthFilteredScan	1	uniform	140.28 ms	59.576 ms	3.266 ms	1.00	0.03	23.86 KB	1.00
TreeWithinDistance	1	uniform	61.26 ms	6.395 ms	0.351 ms	0.44	0.01	116.47 KB	4.88
LengthFilteredScan	2	clustered	202.96 ms	20.580 ms	1.128 ms	1.00	0.01	103.44 KB	1.00
TreeWithinDistance	2	clustered	249.50 ms	12.765 ms	0.700 ms	1.23	0.01	259.28 KB	2.51
LengthFilteredScan	2	uniform	205.87 ms	3.489 ms	0.191 ms	1.00	0.00	54.65 KB	1.00
TreeWithinDistance	2	uniform	215.04 ms	35.860 ms	1.966 ms	1.04	0.01	192.8 KB	3.53
LengthFilteredScan	3	clustered	255.89 ms	82.143 ms	4.503 ms	1.00	0.02	949.9 KB	1.00
TreeWithinDistance	3	clustered	343.08 ms	1.939 ms	0.106 ms	1.34	0.02	1366.63 KB	1.44
LengthFilteredScan	3	uniform	259.01 ms	55.166 ms	3.024 ms	1.00	0.01	741.56 KB	1.00
TreeWithinDistance	3	uniform	326.08 ms	22.624 ms	1.240 ms	1.26	0.01	1152.95 KB	1.55
LengthFilteredScan	4	clustered	294.83 ms	25.630 ms	1.405 ms	1.00	0.01	5113.56 KB	1.00
TreeWithinDistance	4	clustered	395.45 ms	21.577 ms	1.183 ms	1.34	0.01	7216.2 KB	1.41
LengthFilteredScan	4	uniform	304.93 ms	34.331 ms	1.882 ms	1.00	0.01	5514.13 KB	1.00
TreeWithinDistance	4	uniform	396.30 ms	14.014 ms	0.768 ms	1.30	0.01	7964.5 KB	1.44

Lodestar.Text.Benchmarks.BkTreeNearestBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Count	Mean	Error	StdDev	Ratio	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
FullScan	10	76.84 ms	1.942 ms	0.106 ms	1.00	1857.1429	1857.1429	1857.1429	9386.06 KB	1.000
TreeNearest	10	50.33 ms	22.360 ms	1.226 ms	0.66	-	-	-	29.69 KB	0.003
FullScan	2000	76.02 ms	1.965 ms	0.108 ms	1.00	1857.1429	1857.1429	1857.1429	10007.93 KB	1.00
TreeNearest	2000	72.60 ms	8.500 ms	0.466 ms	0.96	-	-	-	1273.29 KB	0.13

Lodestar.Text.Benchmarks.BlockedTableBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	length	Mean	Error	StdDev	Allocated
Latin	1000	44.85 μs	4.238 μs	0.232 μs	-
Cjk	1000	62.23 μs	17.300 μs	0.948 μs	-
Latin	10000	4,119.58 μs	137.029 μs	7.511 μs	-
Cjk	10000	7,884.73 μs	2,034.084 μs	111.495 μs	-
Latin	65536	184,608.86 μs	4,693.602 μs	257.272 μs	-

Lodestar.Text.Benchmarks.BlockedTableBenchmarks-report-github (run 2)

As of 2026-08-27, measured at commit 2483f1a00691271083f00baf3835de96bf0a4076.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.400
  [Host]   : .NET 10.0.11 (10.0.1126.37416), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.11 (10.0.1126.37416), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	length	Mean	Error	StdDev	Allocated
Latin	1000	51.93 μs	0.631 μs	0.035 μs	-
Cjk	1000	56.06 μs	2.402 μs	0.132 μs	-
Latin	10000	5,514.80 μs	1,122.506 μs	61.528 μs	-
Cjk	10000	6,234.71 μs	289.373 μs	15.862 μs	-
Latin	65536	202,152.38 μs	5,826.234 μs	319.356 μs	-

Lodestar.Text.Benchmarks.Bm25Benchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
LodestarQuery	1000	1.587 μs	0.1494 μs	0.0082 μs	1.00	0.01	0.0248	-	-	424 B	1.00
LodestarQueryMultiTerm	1000	2.168 μs	0.3979 μs	0.0218 μs	1.37	0.01	0.0229	-	-	424 B	1.00
LuceneQuery	1000	2.519 μs	0.3591 μs	0.0197 μs	1.59	0.01	0.3128	-	-	5264 B	12.42
LodestarFromText	1000	2,933.817 μs	530.7547 μs	29.0925 μs	1,848.83	17.89	914.0625	890.6250	882.8125	3723446 B	8,781.71
LuceneFromText	1000	5,806.540 μs	309.7757 μs	16.9799 μs	3,659.16	18.76	85.9375	78.1250	7.8125	1375080 B	3,243.11
LodestarQuery	20000	20.922 μs	29.9984 μs	1.6443 μs	1.00	0.10	-	-	-	424 B	1.00
LodestarQueryMultiTerm	20000	29.665 μs	1.0083 μs	0.0553 μs	1.42	0.10	-	-	-	424 B	1.00
LuceneQuery	20000	14.199 μs	0.4799 μs	0.0263 μs	0.68	0.05	0.5035	-	-	8648 B	20.40
LodestarFromText	20000	56,228.574 μs	5,734.7029 μs	314.3384 μs	2,699.14	192.86	1666.6667	1666.6667	1666.6667	61663525 B	145,432.84
LuceneFromText	20000	114,813.689 μs	16,313.0509 μs	894.1734 μs	5,511.40	394.65	1200.0000	1000.0000	-	22417093 B	52,870.50

Lodestar.Text.Benchmarks.BpeBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Unigram	28.22 ms	2.376 ms	0.130 ms	1.00	0.01	500.0000	8.19 MB	1.00
Bpe	84.03 ms	17.995 ms	0.986 ms	2.98	0.03	1666.6667	28.47 MB	3.48

Lodestar.Text.Benchmarks.BpeScalingBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Gen0	Allocated
BpeOnOnePathologicalToken	512	19.04 μs	1.015 μs	0.056 μs	0.4578	7.48 KB
BpeOnOnePathologicalToken	1024	66.55 μs	2.902 μs	0.159 μs	0.8545	14.53 KB
BpeOnOnePathologicalToken	2048	180.25 μs	14.785 μs	0.810 μs	1.7090	28.58 KB
BpeOnOnePathologicalToken	4096	389.09 μs	28.772 μs	1.577 μs	3.4180	56.63 KB

Lodestar.Text.Benchmarks.BpeWordCacheBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  InvocationCount=1  IterationCount=3  
LaunchCount=1  UnrollFactor=1  WarmupCount=3

Method	Mean	Error	StdDev	Allocated
EncodeUnseenProse	9.968 ms	18.156 ms	0.9952 ms	1.7 MB

Lodestar.Text.Benchmarks.BucketRouteDiagnostics-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Alphabet	Mean	Error	StdDev	Allocated
DpGroup	cjk	19.05 μs	2.704 μs	0.148 μs	-
MyersGroup	cjk	241.58 μs	21.001 μs	1.151 μs	-
DpGroup	latin	10.22 μs	3.269 μs	0.179 μs	-
MyersGroup	latin	132.59 μs	15.357 μs	0.842 μs	-

Lodestar.Text.Benchmarks.CdistCutoffBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Cutoff	Widths	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar_Cdist_Bounded	0	False	304.04 μs	145.195 μs	7.959 μs	1.00	0.03	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	0	False	322.01 μs	30.583 μs	1.676 μs	1.06	0.02	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	0	True	335.98 μs	14.661 μs	0.804 μs	1.00	0.00	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	0	True	324.65 μs	78.878 μs	4.324 μs	0.97	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	60	False	320.00 μs	38.113 μs	2.089 μs	1.00	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	60	False	306.68 μs	25.489 μs	1.397 μs	0.96	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	60	True	273.53 μs	12.902 μs	0.707 μs	1.00	0.00	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	60	True	324.63 μs	13.657 μs	0.749 μs	1.19	0.00	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	80	False	305.98 μs	24.946 μs	1.367 μs	1.00	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	80	False	305.68 μs	37.825 μs	2.073 μs	1.00	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	80	True	166.53 μs	15.889 μs	0.871 μs	1.00	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	80	True	324.49 μs	117.965 μs	6.466 μs	1.95	0.03	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	90	False	228.29 μs	14.255 μs	0.781 μs	1.00	0.00	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	90	False	305.10 μs	9.308 μs	0.510 μs	1.34	0.00	1.9531	32.02 KB	1.00
Lodestar_Cdist_Bounded	90	True	89.91 μs	11.856 μs	0.650 μs	1.00	0.01	1.9531	32.02 KB	1.00
Lodestar_Cdist_EveryPair	90	True	323.46 μs	18.948 μs	1.039 μs	3.60	0.02	1.9531	32.02 KB	1.00

Lodestar.Text.Benchmarks.CdistIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Size	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Cdist	50	181.3 μs	56.28 μs	3.08 μs	1.00	0.02	0.9766	-	-	19.55 KB	1.00
Lodestar_HandWrittenLoop	50	169.2 μs	14.54 μs	0.80 μs	0.93	0.01	0.9766	-	-	19.55 KB	1.00
FuzzySharp_ExtractAllPerQuery	50	484.9 μs	35.52 μs	1.95 μs	2.67	0.04	17.5781	-	-	301.95 KB	15.44
Lodestar_Cdist	200	3,251.4 μs	103.37 μs	5.67 μs	1.00	0.00	97.6563	97.6563	97.6563	312.62 KB	1.00
Lodestar_HandWrittenLoop	200	3,205.3 μs	837.64 μs	45.91 μs	0.99	0.01	97.6563	97.6563	97.6563	312.62 KB	1.00
FuzzySharp_ExtractAllPerQuery	200	7,627.8 μs	827.85 μs	45.38 μs	2.35	0.01	289.0625	-	-	4723.45 KB	15.11

Lodestar.Text.Benchmarks.ClassificationReportBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Classes	Mean	Error	StdDev	Gen0	Gen1	Allocated
Report	10	605.0 ns	162.8 ns	8.93 ns	0.1011	-	1.66 KB
Report	1000	213,525.8 ns	19,358.7 ns	1,061.11 ns	7.3242	1.7090	121.55 KB

Lodestar.Text.Benchmarks.ClassifierCurveBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Roc	1000000	149.1 ms	12.09 ms	0.66 ms	750.0000	750.0000	750.0000	53.79 MB
PrecisionRecall	1000000	135.2 ms	21.81 ms	1.20 ms	1000.0000	1000.0000	1000.0000	68.66 MB

Lodestar.Text.Benchmarks.ClusteringAgreementBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Clusters	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
AdjustedRandScore	100000	10	2.081 ms	0.1340 ms	0.0073 ms	-	-	-	12.07 KB
MutualInformationScore	100000	10	2.114 ms	0.1338 ms	0.0073 ms	-	-	-	12.07 KB
AdjustedMutualInformationScore	100000	10	21.352 ms	0.1813 ms	0.0099 ms	218.7500	218.7500	218.7500	1031.09 KB
AdjustedRandScore	100000	100	3.070 ms	1.5101 ms	0.0828 ms	187.5000	187.5000	187.5000	937.61 KB
MutualInformationScore	100000	100	3.159 ms	0.7838 ms	0.0430 ms	210.9375	199.2188	199.2188	937.61 KB
AdjustedMutualInformationScore	100000	100	183.409 ms	9.3612 ms	0.5131 ms	333.3333	333.3333	333.3333	1745.05 KB

Lodestar.Text.Benchmarks.DamerauLevenshteinBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar	12	875.5 ns	232.46 ns	12.74 ns	1.00	0.02	0.0496	-	-	840 B	1.00
F23StringSimilarity	12	1,802.5 ns	45.83 ns	2.51 ns	2.06	0.03	0.0896	-	-	1520 B	1.81
Lodestar	120	68,036.0 ns	5,926.27 ns	324.84 ns	1.00	0.01	0.1221	-	-	2128 B	1.00
F23StringSimilarity	120	151,137.4 ns	5,380.07 ns	294.90 ns	2.22	0.01	3.4180	-	-	61064 B	28.70
Lodestar	1000	4,335,356.7 ns	407,025.52 ns	22,310.44 ns	1.00	0.01	-	-	-	5656 B	1.00
F23StringSimilarity	1000	12,319,400.8 ns	121,617.04 ns	6,666.24 ns	2.84	0.01	984.3750	984.3750	984.3750	4018553 B	710.49

Lodestar.Text.Benchmarks.DbscanDimensionBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Fit	10000x8x8	154.80 ms	66.765 ms	3.660 ms	1.00	0.03	1750.0000	1750.0000	1750.0000	29.42 MB	1.00
NumFlat_Fit	10000x8x8	2,314.73 ms	204.976 ms	11.235 ms	14.96	0.32	14000.0000	8000.0000	7000.0000	156.31 MB	5.31
Lodestar_Fit	5000x16x8	56.51 ms	4.662 ms	0.256 ms	1.00	0.01	1444.4444	1444.4444	1444.4444	13.96 MB	1.00
NumFlat_Fit	5000x16x8	816.82 ms	72.311 ms	3.964 ms	14.45	0.08	5000.0000	3000.0000	3000.0000	59.77 MB	4.28

Lodestar.Text.Benchmarks.DbscanIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Fit	20000x2x10	185.31 ms	10.268 ms	0.563 ms	1.00	0.00	1333.3333	1333.3333	1333.3333	112.66 MB	1.00
NumFlat_Fit	20000x2x10	6,698.39 ms	232.297 ms	12.733 ms	36.15	0.11	38000.0000	14000.0000	11000.0000	531.87 MB	4.72
Dbscan_CalculateClusters	20000x2x10	4,009.57 ms	592.769 ms	32.492 ms	21.64	0.16	22000.0000	12000.0000	10000.0000	372.14 MB	3.30
Lodestar_Fit	5000x2x5	20.72 ms	0.760 ms	0.042 ms	1.00	0.00	1312.5000	1312.5000	1312.5000	14.16 MB	1.00
NumFlat_Fit	5000x2x5	474.80 ms	67.303 ms	3.689 ms	22.91	0.16	4000.0000	3000.0000	2000.0000	66.59 MB	4.70
Dbscan_CalculateClusters	5000x2x5	267.64 ms	28.925 ms	1.586 ms	12.92	0.07	3500.0000	3500.0000	3500.0000	40.4 MB	2.85

Lodestar.Text.Benchmarks.DecompositionBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	RatioSD
TruncatedSvd_Rank20	13.387 ms	2.514 ms	0.1378 ms	1.00	0.01
Nmf_Rank20	109.046 ms	11.975 ms	0.6564 ms	8.15	0.08
Nmf_Transform	2.982 ms	1.079 ms	0.0591 ms	0.22	0.00
MlNet_ProjectToPrincipalComponents_Rank20	24.506 ms	2.182 ms	0.1196 ms	1.83	0.02

Lodestar.Text.Benchmarks.DecompositionWidthBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Columns	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
TruncatedSvd_Rank20	500	20,144.8 μs	2,675.79 μs	146.67 μs	3031.2500	3000.0000	3000.0000	11692.42 KB
TruncatedSvd_Transform	500	538.7 μs	98.61 μs	5.40 μs	90.8203	90.8203	90.8203	312.58 KB
Nmf_Frobenius_Rank20	500	182,884.0 μs	20,189.33 μs	1,106.65 μs	4500.0000	4500.0000	4500.0000	18143.46 KB
Nmf_KullbackLeibler_Rank20	500	255,943.6 μs	60,194.00 μs	3,299.44 μs	4000.0000	4000.0000	4000.0000	18596.18 KB
TruncatedSvd_Rank20	20000	142,515.5 μs	45,258.22 μs	2,480.76 μs	3750.0000	3750.0000	3750.0000	115064.21 KB
TruncatedSvd_Transform	20000	924.0 μs	150.04 μs	8.22 μs	43.9453	43.9453	43.9453	312.55 KB
Nmf_Frobenius_Rank20	20000	715,381.8 μs	132,077.63 μs	7,239.62 μs	5000.0000	5000.0000	5000.0000	166225.79 KB
Nmf_KullbackLeibler_Rank20	20000	576,987.8 μs	31,521.80 μs	1,727.82 μs	5000.0000	5000.0000	5000.0000	166680.8 KB

Lodestar.Text.Benchmarks.DoubleMetaphoneBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Allocated
Encode	282.6 μs	8.02 μs	0.44 μs	29.7852	488.6 KB

Lodestar.Text.Benchmarks.EmbeddingSearchBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Count	Mean	Error	StdDev	Allocated
SearchTop10	10000	525.6 μs	24.75 μs	1.36 μs	1.63 KB
SearchTop10	100000	8,326.0 μs	607.90 μs	33.32 μs	1.63 KB

Lodestar.Text.Benchmarks.EmbeddingSearchHelpersBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
FromBlockNormalized	1000	705.6 μs	236.36 μs	12.96 μs	249.0234	249.0234	249.0234	1500.18 KB
MmrSelect100	1000	7,364.9 μs	111.51 μs	6.11 μs	-	-	-	24.93 KB
FromBlockNormalized	100000	73,221.6 μs	10,908.94 μs	597.96 μs	166.6667	166.6667	166.6667	150002.31 KB
MmrSelect100	100000	7,218.5 μs	85.64 μs	4.69 μs	-	-	-	24.92 KB

Lodestar.Text.Benchmarks.EncoderIncumbentBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]     : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  Job-PVBJKR : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

LaunchCount=1

Method	Job	InvocationCount	IterationCount	RunStrategy	UnrollFactor	WarmupCount	RowCount	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_OneHot	Job-PVBJKR	1	20	Throughput	1	5	1000	280.88 μs	12.920 μs	14.361 μs	1.00	0.07	-	-	-	166.8 KB	1.00
Lodestar_Ordinal	Job-PVBJKR	1	20	Throughput	1	5	1000	280.06 μs	13.464 μs	14.966 μs	1.00	0.07	-	-	-	17.84 KB	0.11
Lodestar_Impute	Job-PVBJKR	1	20	Throughput	1	5	1000	106.92 μs	5.990 μs	6.657 μs	0.38	0.03	-	-	-	32.18 KB	0.19
MlNet_OneHotEncoding_Fit	Job-PVBJKR	1	20	Throughput	1	5	1000	483.38 μs	82.617 μs	95.141 μs	1.73	0.34	-	-	-	34.19 KB	0.20
MlNet_OneHotEncoding_Read	Job-PVBJKR	1	20	Throughput	1	5	1000	1,730.35 μs	142.857 μs	158.785 μs	6.18	0.63	-	-	-	339.34 KB	2.03
MlNet_ReplaceMissingValues_Read	Job-PVBJKR	1	20	Throughput	1	5	1000	1,244.83 μs	57.526 μs	63.940 μs	4.44	0.31	-	-	-	312.66 KB	1.87
Lodestar_OneHot	ShortRun	Default	3	Default	16	3	1000	175.90 μs	20.978 μs	1.150 μs	1.00	0.01	49.8047	49.8047	49.8047	165.86 KB	1.00
Lodestar_Ordinal	ShortRun	Default	3	Default	16	3	1000	70.66 μs	21.747 μs	1.192 μs	0.40	0.01	0.9766	-	-	16.91 KB	0.10
Lodestar_Impute	ShortRun	Default	3	Default	16	3	1000	16.25 μs	0.870 μs	0.048 μs	0.09	0.00	1.8921	-	-	31.42 KB	0.19
MlNet_OneHotEncoding_Fit	ShortRun	Default	3	Default	16	3	1000	955.40 μs	2,552.454 μs	139.909 μs	5.43	0.69	59.5703	7.8125	-	986.36 KB	5.95
MlNet_OneHotEncoding_Read	ShortRun	Default	3	Default	16	3	1000	935.41 μs	338.541 μs	18.557 μs	5.32	0.10	48.8281	7.8125	-	797.9 KB	4.81
MlNet_ReplaceMissingValues_Read	ShortRun	Default	3	Default	16	3	1000	254.35 μs	69.575 μs	3.814 μs	1.45	0.02	19.0430	18.5547	-	299.33 KB	1.80
Lodestar_OneHot	Job-PVBJKR	1	20	Throughput	1	5	20000	5,721.90 μs	861.411 μs	921.701 μs	1.03	0.23	-	-	-	3283.98 KB	1.00
Lodestar_Ordinal	Job-PVBJKR	1	20	Throughput	1	5	20000	4,742.58 μs	853.193 μs	982.538 μs	0.85	0.22	-	-	-	314.71 KB	0.10
Lodestar_Impute	Job-PVBJKR	1	20	Throughput	1	5	20000	690.00 μs	15.139 μs	17.434 μs	0.12	0.02	-	-	-	625.93 KB	0.19
MlNet_OneHotEncoding_Fit	Job-PVBJKR	1	20	Throughput	1	5	20000	4,453.76 μs	43.815 μs	44.994 μs	0.80	0.13	-	-	-	34.19 KB	0.01
MlNet_OneHotEncoding_Read	Job-PVBJKR	1	20	Throughput	1	5	20000	16,567.29 μs	1,268.289 μs	1,302.440 μs	2.97	0.52	-	-	-	445.1 KB	0.14
MlNet_ReplaceMissingValues_Read	Job-PVBJKR	1	20	Throughput	1	5	20000	10,616.54 μs	4,793.414 μs	5,520.103 μs	1.90	1.02	-	-	-	537.01 KB	0.16
Lodestar_OneHot	ShortRun	Default	3	Default	16	3	20000	3,545.25 μs	738.958 μs	40.505 μs	1.00	0.01	992.1875	992.1875	992.1875	3291.64 KB	1.00
Lodestar_Ordinal	ShortRun	Default	3	Default	16	3	20000	3,120.47 μs	307.465 μs	16.853 μs	0.88	0.01	89.8438	89.8438	89.8438	314.42 KB	0.10
Lodestar_Impute	ShortRun	Default	3	Default	16	3	20000	453.33 μs	931.810 μs	51.076 μs	0.13	0.01	128.9063	128.9063	128.9063	626.19 KB	0.19
MlNet_OneHotEncoding_Fit	ShortRun	Default	3	Default	16	3	20000	1,541.66 μs	1,081.474 μs	59.279 μs	0.43	0.02	29.2969	5.8594	-	510.02 KB	0.15
MlNet_OneHotEncoding_Read	ShortRun	Default	3	Default	16	3	20000	4,109.26 μs	435.681 μs	23.881 μs	1.16	0.01	31.2500	15.6250	-	561.76 KB	0.17
MlNet_ReplaceMissingValues_Read	ShortRun	Default	3	Default	16	3	20000	2,396.05 μs	945.286 μs	51.814 μs	0.68	0.01	27.3438	11.7188	-	454.13 KB	0.14

Lodestar.Text.Benchmarks.FilteredVectorSearchBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Records	Mean	Error	StdDev	Gen0	Allocated
FilteredTop10	10000	496.4 μs	294.86 μs	16.16 μs	-	7.17 KB
Unfiltered	10000	563.7 μs	38.83 μs	2.13 μs	-	2.93 KB
HybridSelective	10000	676.0 μs	59.72 μs	3.27 μs	-	15.11 KB
HybridBroad	10000	1,410.1 μs	190.69 μs	10.45 μs	9.7656	342.49 KB
FilteredTop10	100000	9,645.5 μs	2,293.52 μs	125.72 μs	-	7.19 KB
Unfiltered	100000	8,699.8 μs	1,063.07 μs	58.27 μs	-	2.95 KB
HybridSelective	100000	10,071.5 μs	151.19 μs	8.29 μs	-	15.15 KB
HybridBroad	100000	18,105.6 μs	3,159.36 μs	173.18 μs	-	3155.14 KB

Lodestar.Text.Benchmarks.FuzzBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Ratio	97.67 ns	12.53 ns	0.687 ns	1.00	0.01	-	-	NA
PartialRatio	850.15 ns	30.10 ns	1.650 ns	8.70	0.05	-	-	NA
TokenSortRatio	834.71 ns	109.85 ns	6.021 ns	8.55	0.07	0.0668	1120 B	NA
TokenSetRatio	1,004.56 ns	237.78 ns	13.034 ns	10.29	0.13	0.0534	896 B	NA
WRatio	1,292.11 ns	184.58 ns	10.118 ns	13.23	0.12	0.0668	1120 B	NA

Lodestar.Text.Benchmarks.FuzzCodePointBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Allocated	Alloc Ratio
Ratio_RepeatedEmoji	2,029,046.3424 ns	349,000.3044 ns	19,129.8844 ns	1.000	0.01	7.8125	-	160508 B	1.00
TokenSortRatio_EmojiWords	672,436.7337 ns	127,090.2615 ns	6,966.2461 ns	0.331	0.00	35.1563	12.6953	602665 B	3.75
WRatio_ProseAndOneEmoji	1,123,312.3281 ns	313,825.0226 ns	17,201.8085 ns	0.554	0.01	42.9688	15.6250	720569 B	4.49
Ratio_DistinctAstral	70,881,350.7619 ns	3,554,144.2716 ns	194,814.6412 ns	34.935	0.30	-	-	510817 B	3.18
WRatio_EmptyOperand	0.0000 ns	0.0000 ns	0.0000 ns	0.000	0.00	-	-	-	0.00

Lodestar.Text.Benchmarks.FuzzIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Operation	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar	Ratio	98.49 ns	9.530 ns	0.522 ns	1.00	0.01	-	-	NA
FuzzySharp	Ratio	208.71 ns	31.552 ns	1.729 ns	2.12	0.02	0.0048	80 B	NA
Lodestar	PartialRatio	753.21 ns	51.939 ns	2.847 ns	1.00	0.00	-	-	NA
FuzzySharp	PartialRatio	10,565.73 ns	934.799 ns	51.240 ns	14.03	0.07	-	160 B	NA
Lodestar	TokenSetRatio	1,029.99 ns	341.266 ns	18.706 ns	1.00	0.02	0.0534	896 B	1.00
FuzzySharp	TokenSetRatio	2,130.73 ns	385.982 ns	21.157 ns	2.07	0.04	0.1144	1944 B	2.17
Lodestar	WRatio	1,273.40 ns	184.778 ns	10.128 ns	1.00	0.01	0.0668	1120 B	1.00
FuzzySharp	WRatio	5,035.12 ns	250.338 ns	13.722 ns	3.95	0.03	0.1831	3096 B	2.76

Lodestar.Text.Benchmarks.HouseholderQrBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Householder	2000	1.574 ms	0.1505 ms	0.0082 ms	277.3438	267.5781	250.0000	1.38 MB
Householder	20000	18.539 ms	2.6219 ms	0.1437 ms	968.7500	968.7500	968.7500	13.74 MB

Lodestar.Text.Benchmarks.IndelBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_Utf16	8	26.61 ns	2.500 ns	0.137 ns	1.00	0.01	-	NA
Distance_CodePoint	8	40.87 ns	0.064 ns	0.003 ns	1.54	0.01	-	NA
NormalizedSimilarity_Utf16	8	27.51 ns	2.435 ns	0.133 ns	1.03	0.01	-	NA
SubsequenceLength_Utf16	8	26.31 ns	2.699 ns	0.148 ns	0.99	0.01	-	NA
Distance_Utf16	12	28.17 ns	3.132 ns	0.172 ns	1.00	0.01	-	NA
Distance_CodePoint	12	42.37 ns	5.950 ns	0.326 ns	1.50	0.01	-	NA
NormalizedSimilarity_Utf16	12	28.81 ns	0.852 ns	0.047 ns	1.02	0.01	-	NA
SubsequenceLength_Utf16	12	28.29 ns	3.782 ns	0.207 ns	1.00	0.01	-	NA
Distance_Utf16	16	30.57 ns	3.172 ns	0.174 ns	1.00	0.01	-	NA
Distance_CodePoint	16	46.50 ns	2.113 ns	0.116 ns	1.52	0.01	-	NA
NormalizedSimilarity_Utf16	16	31.69 ns	1.603 ns	0.088 ns	1.04	0.01	-	NA
SubsequenceLength_Utf16	16	30.93 ns	10.858 ns	0.595 ns	1.01	0.02	-	NA
Distance_Utf16	20	35.42 ns	1.110 ns	0.061 ns	1.00	0.00	-	NA
Distance_CodePoint	20	46.23 ns	3.283 ns	0.180 ns	1.31	0.00	-	NA
NormalizedSimilarity_Utf16	20	36.58 ns	3.845 ns	0.211 ns	1.03	0.01	-	NA
SubsequenceLength_Utf16	20	35.02 ns	14.755 ns	0.809 ns	0.99	0.02	-	NA
Distance_Utf16	24	69.76 ns	32.844 ns	1.800 ns	1.00	0.03	-	NA
Distance_CodePoint	24	61.47 ns	1.301 ns	0.071 ns	0.88	0.02	-	NA
NormalizedSimilarity_Utf16	24	57.02 ns	0.424 ns	0.023 ns	0.82	0.02	-	NA
SubsequenceLength_Utf16	24	56.28 ns	4.206 ns	0.231 ns	0.81	0.02	-	NA
Distance_Utf16	32	182.17 ns	56.914 ns	3.120 ns	1.00	0.02	-	NA
Distance_CodePoint	32	69.09 ns	1.329 ns	0.073 ns	0.38	0.01	-	NA
NormalizedSimilarity_Utf16	32	66.09 ns	0.208 ns	0.011 ns	0.36	0.01	-	NA
SubsequenceLength_Utf16	32	62.62 ns	1.674 ns	0.092 ns	0.34	0.01	-	NA
Distance_Utf16	128	378.71 ns	10.085 ns	0.553 ns	1.00	0.00	-	NA
Distance_CodePoint	128	387.84 ns	70.776 ns	3.879 ns	1.02	0.01	-	NA
NormalizedSimilarity_Utf16	128	381.98 ns	108.734 ns	5.960 ns	1.01	0.01	-	NA
SubsequenceLength_Utf16	128	375.05 ns	5.447 ns	0.299 ns	0.99	0.00	-	NA
Distance_Utf16	512	4,562.39 ns	180.380 ns	9.887 ns	1.00	0.00	-	NA
Distance_CodePoint	512	4,777.97 ns	73.017 ns	4.002 ns	1.05	0.00	-	NA
NormalizedSimilarity_Utf16	512	4,665.94 ns	860.018 ns	47.140 ns	1.02	0.01	-	NA
SubsequenceLength_Utf16	512	4,649.02 ns	947.274 ns	51.923 ns	1.02	0.01	-	NA

Lodestar.Text.Benchmarks.IndelCodePointBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_CodePoint	20	386.2 ns	15.12 ns	0.83 ns	1.00	0.00	-	NA
Distance_Utf16	20	197.1 ns	136.13 ns	7.46 ns	0.51	0.02	-	NA
Distance_CodePoint	128	2,399.4 ns	426.74 ns	23.39 ns	1.00	0.01	-	NA
Distance_Utf16	128	3,049.4 ns	162.53 ns	8.91 ns	1.27	0.01	-	NA
Distance_CodePoint	512	13,470.1 ns	431.16 ns	23.63 ns	1.00	0.00	-	NA
Distance_Utf16	512	32,552.8 ns	11,011.35 ns	603.57 ns	2.42	0.04	-	NA

Lodestar.Text.Benchmarks.IndelPatternBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Corpus	Length	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar_Pairwise	beyondTable	4	8,719.1 ns	3,331.66 ns	182.62 ns	1.00	0.03	-	-	NA
Lodestar_HeldPattern	beyondTable	4	8,091.2 ns	4,531.95 ns	248.41 ns	0.93	0.03	-	40 B	NA
FuzzySharp_PerText	beyondTable	4	11,763.0 ns	3,019.31 ns	165.50 ns	1.35	0.03	0.3052	5120 B	NA
Lodestar_Pairwise	beyondTable	16	10,110.7 ns	1,810.90 ns	99.26 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	beyondTable	16	9,355.4 ns	1,620.46 ns	88.82 ns	0.93	0.01	-	40 B	NA
FuzzySharp_PerText	beyondTable	16	14,591.8 ns	1,905.94 ns	104.47 ns	1.44	0.02	0.3052	5120 B	NA
Lodestar_Pairwise	beyondTable	32	12,788.0 ns	4,136.73 ns	226.75 ns	1.00	0.02	-	-	NA
Lodestar_HeldPattern	beyondTable	32	11,280.6 ns	114.90 ns	6.30 ns	0.88	0.01	-	40 B	NA
FuzzySharp_PerText	beyondTable	32	18,041.3 ns	1,807.43 ns	99.07 ns	1.41	0.02	0.3052	5120 B	NA
Lodestar_Pairwise	longAffix	4	5,608.8 ns	3,506.08 ns	192.18 ns	1.00	0.04	-	-	NA
Lodestar_HeldPattern	longAffix	4	6,413.3 ns	1,307.82 ns	71.69 ns	1.14	0.04	-	40 B	NA
FuzzySharp_PerText	longAffix	4	9,344.8 ns	1,572.37 ns	86.19 ns	1.67	0.05	0.3052	5120 B	NA
Lodestar_Pairwise	longAffix	16	6,901.5 ns	3,550.69 ns	194.63 ns	1.00	0.03	-	-	NA
Lodestar_HeldPattern	longAffix	16	6,909.1 ns	507.73 ns	27.83 ns	1.00	0.02	-	40 B	NA
FuzzySharp_PerText	longAffix	16	11,323.6 ns	1,160.36 ns	63.60 ns	1.64	0.04	0.3052	5120 B	NA
Lodestar_Pairwise	longAffix	32	9,202.8 ns	3,486.55 ns	191.11 ns	1.00	0.03	-	-	NA
Lodestar_HeldPattern	longAffix	32	8,989.0 ns	6,085.76 ns	333.58 ns	0.98	0.04	-	40 B	NA
FuzzySharp_PerText	longAffix	32	15,051.5 ns	955.24 ns	52.36 ns	1.64	0.03	0.3052	5120 B	NA
Lodestar_Pairwise	oneWordAffix	4	3,728.5 ns	417.05 ns	22.86 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	oneWordAffix	4	4,082.2 ns	390.97 ns	21.43 ns	1.09	0.01	-	40 B	NA
FuzzySharp_PerText	oneWordAffix	4	7,454.3 ns	463.22 ns	25.39 ns	2.00	0.01	0.3052	5120 B	NA
Lodestar_Pairwise	oneWordAffix	16	5,344.9 ns	655.12 ns	35.91 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	oneWordAffix	16	5,358.8 ns	1,344.58 ns	73.70 ns	1.00	0.01	-	40 B	NA
FuzzySharp_PerText	oneWordAffix	16	9,973.0 ns	810.46 ns	44.42 ns	1.87	0.01	0.3052	5120 B	NA
Lodestar_Pairwise	oneWordAffix	32	7,780.6 ns	1,307.89 ns	71.69 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	oneWordAffix	32	7,364.1 ns	502.14 ns	27.52 ns	0.95	0.01	-	40 B	NA
FuzzySharp_PerText	oneWordAffix	32	13,813.3 ns	1,868.26 ns	102.41 ns	1.78	0.02	0.3052	5120 B	NA
Lodestar_Pairwise	sharing24	4	2,373.6 ns	1,269.23 ns	69.57 ns	1.00	0.04	-	-	NA
Lodestar_HeldPattern	sharing24	4	2,129.4 ns	221.52 ns	12.14 ns	0.90	0.02	-	40 B	NA
FuzzySharp_PerText	sharing24	4	5,911.2 ns	493.34 ns	27.04 ns	2.49	0.06	0.3052	5120 B	NA
Lodestar_Pairwise	sharing24	16	4,040.4 ns	1,155.54 ns	63.34 ns	1.00	0.02	-	-	NA
Lodestar_HeldPattern	sharing24	16	2,871.7 ns	206.66 ns	11.33 ns	0.71	0.01	-	40 B	NA
FuzzySharp_PerText	sharing24	16	8,492.4 ns	623.47 ns	34.17 ns	2.10	0.03	0.3052	5120 B	NA
Lodestar_Pairwise	sharing24	32	6,577.2 ns	574.27 ns	31.48 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	sharing24	32	3,800.8 ns	395.86 ns	21.70 ns	0.58	0.00	-	40 B	NA
FuzzySharp_PerText	sharing24	32	12,639.1 ns	610.80 ns	33.48 ns	1.92	0.01	0.3052	5120 B	NA
Lodestar_Pairwise	suffixOnly	4	3,659.3 ns	412.93 ns	22.63 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	suffixOnly	4	3,383.9 ns	268.46 ns	14.72 ns	0.92	0.01	-	40 B	NA
FuzzySharp_PerText	suffixOnly	4	7,206.2 ns	1,169.11 ns	64.08 ns	1.97	0.02	0.3052	5120 B	NA
Lodestar_Pairwise	suffixOnly	16	5,214.1 ns	840.69 ns	46.08 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	suffixOnly	16	4,098.6 ns	187.19 ns	10.26 ns	0.79	0.01	-	40 B	NA
FuzzySharp_PerText	suffixOnly	16	10,346.3 ns	2,603.94 ns	142.73 ns	1.98	0.03	0.3052	5120 B	NA
Lodestar_Pairwise	suffixOnly	32	7,761.2 ns	493.36 ns	27.04 ns	1.00	0.00	-	-	NA
Lodestar_HeldPattern	suffixOnly	32	8,306.8 ns	304.01 ns	16.66 ns	1.07	0.00	-	40 B	NA
FuzzySharp_PerText	suffixOnly	32	14,003.7 ns	500.06 ns	27.41 ns	1.80	0.01	0.3052	5120 B	NA
Lodestar_Pairwise	unrelated	4	1,440.7 ns	89.58 ns	4.91 ns	1.00	0.00	-	-	NA
Lodestar_HeldPattern	unrelated	4	571.6 ns	34.84 ns	1.91 ns	0.40	0.00	0.0019	40 B	NA
FuzzySharp_PerText	unrelated	4	5,115.3 ns	243.20 ns	13.33 ns	3.55	0.01	0.3052	5120 B	NA
Lodestar_Pairwise	unrelated	16	2,933.9 ns	364.09 ns	19.96 ns	1.00	0.01	-	-	NA
Lodestar_HeldPattern	unrelated	16	1,395.7 ns	249.92 ns	13.70 ns	0.48	0.00	0.0019	40 B	NA
FuzzySharp_PerText	unrelated	16	7,635.7 ns	737.49 ns	40.42 ns	2.60	0.02	0.3052	5120 B	NA
Lodestar_Pairwise	unrelated	32	5,253.7 ns	145.46 ns	7.97 ns	1.00	0.00	-	-	NA
Lodestar_HeldPattern	unrelated	32	2,355.7 ns	452.49 ns	24.80 ns	0.45	0.00	-	40 B	NA
FuzzySharp_PerText	unrelated	32	11,889.3 ns	998.41 ns	54.73 ns	2.26	0.01	0.3052	5120 B	NA

Lodestar.Text.Benchmarks.KMeansFitIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Fit	10000x16x16	5,523.6 μs	297.77 μs	16.32 μs	1.00	0.00	250.0000	242.1875	242.1875	1495.92 KB	1.000
NumFlat_Fit	10000x16x16	39,789.7 μs	22,604.40 μs	1,239.02 μs	7.20	0.20	-	-	-	10.39 KB	0.007
MetaNumerics_Fit	10000x16x16	84,565.1 μs	6,199.92 μs	339.84 μs	15.31	0.07	-	-	-	41.95 KB	0.028
Lodestar_Fit	10000x2x8	806.8 μs	56.63 μs	3.10 μs	1.00	0.00	49.8047	49.8047	49.8047	391.94 KB	1.000
NumFlat_Fit	10000x2x8	4,796.1 μs	145.83 μs	7.99 μs	5.94	0.02	-	-	-	3 KB	0.008
MetaNumerics_Fit	10000x2x8	3,041.8 μs	172.30 μs	9.44 μs	3.77	0.02	-	-	-	39.57 KB	0.101
Lodestar_Fit	50000x8x32	27,358.2 μs	4,236.95 μs	232.24 μs	1.00	0.01	343.7500	343.7500	343.7500	4309.02 KB	1.000
NumFlat_Fit	50000x8x32	451,374.8 μs	125,566.18 μs	6,882.71 μs	16.50	0.25	-	-	-	16.29 KB	0.004
MetaNumerics_Fit	50000x8x32	755,246.8 μs	173,090.21 μs	9,487.66 μs	27.61	0.36	-	-	-	199.91 KB	0.046

Lodestar.Text.Benchmarks.KMeansLloydIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Lloyd	10000x16x16	4.912 ms	0.4225 ms	0.0232 ms	1.00	0.01	250.0000	242.1875	242.1875	1426.34 KB	1.000
NumFlat_Lloyd	10000x16x16	16.623 ms	0.2190 ms	0.0120 ms	3.38	0.01	-	-	-	13.6 KB	0.010
Lodestar_Lloyd	10000x2x8	25.779 ms	3.4786 ms	0.1907 ms	1.00	0.01	31.2500	31.2500	31.2500	360.4 KB	1.00
NumFlat_Lloyd	10000x2x8	114.504 ms	10.9973 ms	0.6028 ms	4.44	0.03	-	-	-	96.47 KB	0.27
Lodestar_Lloyd	50000x8x32	48.205 ms	8.9061 ms	0.4882 ms	1.00	0.01	181.8182	181.8182	181.8182	3946.72 KB	1.000
NumFlat_Lloyd	50000x8x32	207.116 ms	23.7249 ms	1.3004 ms	4.30	0.04	-	-	-	36.47 KB	0.009

Lodestar.Text.Benchmarks.LcsGateBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Band	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Dp	8	131.60 ns	27.070 ns	1.484 ns	1.00	0.01	-	NA
Kernel	8	56.24 ns	11.950 ns	0.655 ns	0.43	0.01	-	NA
Dp_Cjk	8	130.64 ns	3.254 ns	0.178 ns	0.99	0.01	-	NA
Kernel_Cjk	8	98.49 ns	8.541 ns	0.468 ns	0.75	0.01	-	NA
Dp	12	219.70 ns	86.348 ns	4.733 ns	1.00	0.03	-	NA
Kernel	12	62.21 ns	0.898 ns	0.049 ns	0.28	0.01	-	NA
Dp_Cjk	12	224.80 ns	146.443 ns	8.027 ns	1.02	0.04	-	NA
Kernel_Cjk	12	109.46 ns	3.930 ns	0.215 ns	0.50	0.01	-	NA
Dp	14	282.65 ns	152.343 ns	8.350 ns	1.00	0.04	-	NA
Kernel	14	68.54 ns	15.898 ns	0.871 ns	0.24	0.01	-	NA
Dp_Cjk	14	280.19 ns	375.319 ns	20.572 ns	0.99	0.07	-	NA
Kernel_Cjk	14	115.36 ns	3.270 ns	0.179 ns	0.41	0.01	-	NA
Dp	16	370.47 ns	297.662 ns	16.316 ns	1.00	0.05	-	NA
Kernel	16	72.37 ns	2.516 ns	0.138 ns	0.20	0.01	-	NA
Dp_Cjk	16	347.40 ns	6.113 ns	0.335 ns	0.94	0.04	-	NA
Kernel_Cjk	16	120.31 ns	21.754 ns	1.192 ns	0.33	0.01	-	NA
Dp	18	436.48 ns	11.979 ns	0.657 ns	1.00	0.00	-	NA
Kernel	18	75.29 ns	13.850 ns	0.759 ns	0.17	0.00	-	NA
Dp_Cjk	18	446.05 ns	153.069 ns	8.390 ns	1.02	0.02	-	NA
Kernel_Cjk	18	127.29 ns	59.586 ns	3.266 ns	0.29	0.01	-	NA
Dp	20	779.00 ns	202.601 ns	11.105 ns	1.00	0.02	-	NA
Kernel	20	81.06 ns	0.707 ns	0.039 ns	0.10	0.00	-	NA
Dp_Cjk	20	772.27 ns	6.188 ns	0.339 ns	0.99	0.01	-	NA
Kernel_Cjk	20	142.14 ns	0.971 ns	0.053 ns	0.18	0.00	-	NA
Dp	24	996.33 ns	56.884 ns	3.118 ns	1.00	0.00	-	NA
Kernel	24	88.37 ns	2.385 ns	0.131 ns	0.09	0.00	-	NA
Dp_Cjk	24	972.06 ns	175.368 ns	9.613 ns	0.98	0.01	-	NA
Kernel_Cjk	24	141.61 ns	7.956 ns	0.436 ns	0.14	0.00	-	NA
Dp	32	1,536.44 ns	338.577 ns	18.559 ns	1.00	0.01	-	NA
Kernel	32	105.30 ns	0.176 ns	0.010 ns	0.07	0.00	-	NA
Dp_Cjk	32	1,566.11 ns	123.704 ns	6.781 ns	1.02	0.01	-	NA
Kernel_Cjk	32	167.72 ns	31.233 ns	1.712 ns	0.11	0.00	-	NA
Dp	48	3,196.80 ns	545.600 ns	29.906 ns	1.00	0.01	-	NA
Kernel	48	133.94 ns	25.720 ns	1.410 ns	0.04	0.00	-	NA
Dp_Cjk	48	3,259.42 ns	370.936 ns	20.332 ns	1.02	0.01	-	NA
Kernel_Cjk	48	241.43 ns	7.346 ns	0.403 ns	0.08	0.00	-	NA
Dp	64	6,186.52 ns	3,846.784 ns	210.855 ns	1.00	0.04	-	NA
Kernel	64	165.36 ns	22.865 ns	1.253 ns	0.03	0.00	-	NA
Dp_Cjk	64	5,157.25 ns	1,233.798 ns	67.629 ns	0.83	0.03	-	NA
Kernel_Cjk	64	296.20 ns	101.831 ns	5.582 ns	0.05	0.00	-	NA
Dp	96	13,518.92 ns	9,688.116 ns	531.038 ns	1.00	0.05	-	NA
Kernel	96	320.65 ns	7.354 ns	0.403 ns	0.02	0.00	-	NA
Dp_Cjk	96	13,733.97 ns	5,961.090 ns	326.747 ns	1.02	0.04	-	NA
Kernel_Cjk	96	927.41 ns	116.981 ns	6.412 ns	0.07	0.00	-	NA

Lodestar.Text.Benchmarks.LevenshteinBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_Utf16	8	24.87 ns	2.770 ns	0.152 ns	1.00	0.01	-	NA
Distance_Utf16_Cjk	8	24.75 ns	0.230 ns	0.013 ns	1.00	0.01	-	NA
Distance_CodePoint	8	120.11 ns	40.662 ns	2.229 ns	4.83	0.08	-	NA
NormalizedSimilarity_Utf16	8	25.30 ns	5.074 ns	0.278 ns	1.02	0.01	-	NA
Distance_Utf16	64	278.92 ns	76.000 ns	4.166 ns	1.00	0.02	-	NA
Distance_Utf16_Cjk	64	363.31 ns	35.079 ns	1.923 ns	1.30	0.02	-	NA
Distance_CodePoint	64	659.74 ns	20.275 ns	1.111 ns	2.37	0.03	-	NA
NormalizedSimilarity_Utf16	64	278.07 ns	2.019 ns	0.111 ns	1.00	0.01	-	NA
Distance_Utf16	128	895.61 ns	15.406 ns	0.844 ns	1.00	0.00	-	NA
Distance_Utf16_Cjk	128	1,896.21 ns	214.417 ns	11.753 ns	2.12	0.01	-	NA
Distance_CodePoint	128	1,628.58 ns	163.740 ns	8.975 ns	1.82	0.01	-	NA
NormalizedSimilarity_Utf16	128	911.33 ns	17.316 ns	0.949 ns	1.02	0.00	-	NA
Distance_Utf16	512	11,151.84 ns	676.012 ns	37.054 ns	1.00	0.00	-	NA
Distance_Utf16_Cjk	512	16,781.80 ns	1,740.843 ns	95.422 ns	1.50	0.01	-	NA
Distance_CodePoint	512	13,480.59 ns	625.235 ns	34.271 ns	1.21	0.00	-	NA
NormalizedSimilarity_Utf16	512	10,941.47 ns	89.284 ns	4.894 ns	0.98	0.00	-	NA

Lodestar.Text.Benchmarks.LevenshteinCodePointBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Distinct	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_CodePoint	16	32	342.4 ns	79.95 ns	4.38 ns	1.00	0.02	-	NA
Distance_Utf16	16	32	237.1 ns	22.23 ns	1.22 ns	0.69	0.01	-	NA
Distance_CodePoint	16	512	332.3 ns	38.28 ns	2.10 ns	1.00	0.01	-	NA
Distance_Utf16	16	512	243.6 ns	26.15 ns	1.43 ns	0.73	0.01	-	NA
Distance_CodePoint	24	32	427.4 ns	23.41 ns	1.28 ns	1.00	0.00	-	NA
Distance_Utf16	24	32	321.3 ns	25.91 ns	1.42 ns	0.75	0.00	-	NA
Distance_CodePoint	24	512	416.3 ns	75.48 ns	4.14 ns	1.00	0.01	-	NA
Distance_Utf16	24	512	323.4 ns	12.45 ns	0.68 ns	0.78	0.01	-	NA
Distance_CodePoint	32	32	512.0 ns	97.51 ns	5.34 ns	1.00	0.01	-	NA
Distance_Utf16	32	32	414.9 ns	58.88 ns	3.23 ns	0.81	0.01	-	NA
Distance_CodePoint	32	512	525.3 ns	9.99 ns	0.55 ns	1.00	0.00	-	NA
Distance_Utf16	32	512	416.2 ns	7.16 ns	0.39 ns	0.79	0.00	-	NA
Distance_CodePoint	40	32	603.1 ns	184.30 ns	10.10 ns	1.00	0.02	-	NA
Distance_Utf16	40	32	1,256.2 ns	70.06 ns	3.84 ns	2.08	0.03	-	NA
Distance_CodePoint	40	512	590.6 ns	183.85 ns	10.08 ns	1.00	0.02	-	NA
Distance_Utf16	40	512	1,277.9 ns	30.71 ns	1.68 ns	2.16	0.03	-	NA
Distance_CodePoint	128	32	1,940.8 ns	329.84 ns	18.08 ns	1.00	0.01	-	NA
Distance_Utf16	128	32	5,562.9 ns	437.36 ns	23.97 ns	2.87	0.03	-	NA
Distance_CodePoint	128	512	1,932.9 ns	294.13 ns	16.12 ns	1.00	0.01	-	NA
Distance_Utf16	128	512	5,424.9 ns	252.83 ns	13.86 ns	2.81	0.02	-	NA
Distance_CodePoint	512	32	14,727.7 ns	497.40 ns	27.26 ns	1.00	0.00	-	NA
Distance_Utf16	512	32	63,966.6 ns	3,953.18 ns	216.69 ns	4.34	0.01	-	NA
Distance_CodePoint	512	512	451,807.7 ns	194,471.60 ns	10,659.64 ns	1.00	0.03	-	NA
Distance_Utf16	512	512	61,180.4 ns	1,902.82 ns	104.30 ns	0.14	0.00	-	NA

Lodestar.Text.Benchmarks.LevenshteinIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar	8	25.01 ns	5.908 ns	0.324 ns	1.00	0.02	-	-	NA
Fastenshtein	8	85.11 ns	43.865 ns	2.404 ns	3.40	0.09	0.0033	56 B	NA
Quickenshtein	8	85.49 ns	9.493 ns	0.520 ns	3.42	0.04	-	-	NA
F23_StringSimilarity	8	178.69 ns	11.974 ns	0.656 ns	7.15	0.08	0.0076	128 B	NA
Lodestar	64	276.48 ns	23.559 ns	1.291 ns	1.00	0.01	-	-	NA
Fastenshtein	64	5,804.92 ns	6,769.100 ns	371.037 ns	21.00	1.17	0.0153	280 B	NA
Quickenshtein	64	1,373.20 ns	477.138 ns	26.154 ns	4.97	0.08	-	-	NA
F23_StringSimilarity	64	10,620.62 ns	2,108.984 ns	115.601 ns	38.41	0.39	0.0305	576 B	NA
Lodestar	512	11,020.42 ns	354.182 ns	19.414 ns	1.00	0.00	-	-	NA
Fastenshtein	512	391,386.07 ns	75,631.651 ns	4,145.626 ns	35.51	0.33	-	2072 B	NA
Quickenshtein	512	38,861.51 ns	5,220.594 ns	286.158 ns	3.53	0.02	-	-	NA
F23_StringSimilarity	512	705,503.88 ns	60,921.726 ns	3,339.325 ns	64.02	0.28	-	4161 B	NA

Lodestar.Text.Benchmarks.ManyLabelsMetricsBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Labels	Mean	Error	StdDev	Gen0	Gen1	Allocated
F1Macro	1000	602.8 μs	37.10 μs	2.03 μs	3.9063	-	67.71 KB
PrecisionWeighted	1000	605.5 μs	91.74 μs	5.03 μs	3.9063	-	67.71 KB
ReportRows	1000	601.7 μs	57.11 μs	3.13 μs	8.7891	1.9531	153.99 KB
F1Macro	5000	769.9 μs	53.57 μs	2.94 μs	20.5078	5.8594	337.24 KB
PrecisionWeighted	5000	801.7 μs	27.00 μs	1.48 μs	20.5078	5.8594	337.24 KB
ReportRows	5000	918.0 μs	6.98 μs	0.38 μs	46.8750	18.5547	767.27 KB

Lodestar.Text.Benchmarks.MetaNumericsPcaBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_ExplainedVarianceRatio	2000x10	127.93 μs	3.019 μs	0.165 μs	1.00	0.00	-	-	-	2.45 KB	1.00
MetaNumerics_VarianceFraction	2000x10	164,841.36 μs	62,701.689 μs	3,436.891 μs	1,288.52	23.31	750.0000	750.0000	750.0000	31413.75 KB	12,805.60
Lodestar_ExplainedVarianceRatio	2000x50	2,791.14 μs	216.094 μs	11.845 μs	1.00	0.01	-	-	-	42.14 KB	1.00
MetaNumerics_VarianceFraction	2000x50	758,740.96 μs	46,345.098 μs	2,540.331 μs	271.84	1.27	-	-	-	32054.48 KB	760.60
Lodestar_ExplainedVarianceRatio	200x10	22.46 μs	1.080 μs	0.059 μs	1.00	0.00	0.1221	-	-	2.45 KB	1.00
MetaNumerics_VarianceFraction	200x10	891.45 μs	32.906 μs	1.804 μs	39.70	0.11	99.6094	99.6094	99.6094	329.71 KB	134.40

Lodestar.Text.Benchmarks.MetricsBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Classes	Mean	Error	StdDev	Gen0	Allocated
Matrix	1000	2	4,564.7 ns	2,551.85 ns	139.88 ns	0.0229	400 B
MatrixWeighted	1000	2	7,273.7 ns	798.04 ns	43.74 ns	0.0229	400 B
AccuracyScore	1000	2	110.6 ns	25.91 ns	1.42 ns	-	-
F1Macro	1000	2	4,420.1 ns	680.96 ns	37.33 ns	0.0381	656 B
Report	1000	2	7,460.5 ns	216.64 ns	11.88 ns	0.3204	5464 B
Matrix	1000	10	5,056.2 ns	515.11 ns	28.24 ns	0.0763	1336 B
MatrixWeighted	1000	10	7,452.3 ns	511.31 ns	28.03 ns	0.0763	1336 B
AccuracyScore	1000	10	110.9 ns	63.25 ns	3.47 ns	-	-
F1Macro	1000	10	5,325.7 ns	911.33 ns	49.95 ns	0.1068	1880 B
Report	1000	10	11,776.2 ns	964.29 ns	52.86 ns	0.7324	12488 B
Matrix	100000	2	461,028.0 ns	2,528.44 ns	138.59 ns	-	400 B
MatrixWeighted	100000	2	839,140.1 ns	176,586.12 ns	9,679.28 ns	-	401 B
AccuracyScore	100000	2	12,539.4 ns	1,101.11 ns	60.36 ns	-	-
F1Macro	100000	2	460,963.5 ns	29,809.38 ns	1,633.95 ns	-	656 B
Report	100000	2	482,775.4 ns	103,721.23 ns	5,685.31 ns	-	5489 B
Matrix	100000	10	507,442.7 ns	55,749.63 ns	3,055.83 ns	-	1337 B
MatrixWeighted	100000	10	944,366.6 ns	27,360.27 ns	1,499.71 ns	-	1337 B
AccuracyScore	100000	10	12,150.3 ns	796.54 ns	43.66 ns	-	-
F1Macro	100000	10	494,328.4 ns	12,887.64 ns	706.42 ns	-	1881 B
Report	100000	10	501,865.8 ns	47,912.71 ns	2,626.26 ns	-	12833 B
Matrix	1000000	2	4,761,677.0 ns	35,181.97 ns	1,928.44 ns	-	406 B
MatrixWeighted	1000000	2	8,360,905.7 ns	225,109.12 ns	12,338.99 ns	-	412 B
AccuracyScore	1000000	2	120,964.9 ns	6,765.07 ns	370.82 ns	-	-
F1Macro	1000000	2	4,598,766.5 ns	116,150.50 ns	6,366.60 ns	-	662 B
Report	1000000	2	4,635,543.7 ns	296,188.51 ns	16,235.09 ns	-	5510 B
Matrix	1000000	10	4,923,088.3 ns	84,846.61 ns	4,650.73 ns	-	1342 B
MatrixWeighted	1000000	10	9,395,241.4 ns	1,476,496.74 ns	80,931.77 ns	-	1348 B
AccuracyScore	1000000	10	120,923.5 ns	11,063.55 ns	606.43 ns	-	-
F1Macro	1000000	10	4,938,611.4 ns	402,608.45 ns	22,068.33 ns	-	1886 B
Report	1000000	10	5,109,723.8 ns	717,743.75 ns	39,341.96 ns	-	12878 B

Lodestar.Text.Benchmarks.MetricsIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Request	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar	100000	Bundle	10,223.77 μs	425.607 μs	23.329 μs	1.00	0.00	-	-	-	1296 B	1.00
MlNet	100000	Bundle	34,881.92 μs	2,306.082 μs	126.404 μs	3.41	0.01	600.0000	600.0000	600.0000	5090157 B	3,927.59
Lodestar	100000	AccuracyAlone	11.17 μs	1.796 μs	0.098 μs	1.00	0.01	-	-	-	-	NA
MlNet	100000	AccuracyAlone	35,022.17 μs	7,723.659 μs	423.360 μs	3,136.74	40.70	538.4615	538.4615	538.4615	5089463 B	NA
Lodestar	1000000	Bundle	143,944.08 μs	37,372.673 μs	2,048.522 μs	1.00	0.02	-	-	-	-	NA
MlNet	1000000	Bundle	224,334.49 μs	20,373.476 μs	1,116.739 μs	1.56	0.02	-	-	-	23228008 B	NA
Lodestar	1000000	AccuracyAlone	136.25 μs	15.668 μs	0.859 μs	1.00	0.01	-	-	-	-	NA
MlNet	1000000	AccuracyAlone	224,526.36 μs	7,966.915 μs	436.694 μs	1,647.90	9.39	-	-	-	23228008 B	NA

Lodestar.Text.Benchmarks.MultiClassRocAucBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Mean	Error	StdDev	Allocated
OneVsRest	100000	40.63 ms	17.79 ms	0.975 ms	-
OneVsOne	100000	96.33 ms	15.37 ms	0.843 ms	803600 B
OneVsRest	1000000	744.39 ms	1,484.84 ms	81.389 ms	-
OneVsOne	1000000	1,332.74 ms	325.45 ms	17.839 ms	8005264 B

Lodestar.Text.Benchmarks.MultilabelConfusionMatrixBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
PerLabel	100000	6.822 ms	0.1232 ms	0.0068 ms	-	-	-	7.84 KB
PerLabelWeighted	100000	11.475 ms	0.3531 ms	0.0194 ms	-	-	-	7.85 KB
PerSample	100000	72.602 ms	60.2367 ms	3.3018 ms	3142.8571	3000.0000	857.1429	39063.2 KB
PerClass	100000	4.617 ms	0.1025 ms	0.0056 ms	-	-	-	7.99 KB

Lodestar.Text.Benchmarks.MyersGateBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Band	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Dp	4	73.99 ns	1.968 ns	0.108 ns	1.00	0.00	-	NA
Kernel	4	73.99 ns	15.210 ns	0.834 ns	1.00	0.01	-	NA
Dp_Cjk	4	75.04 ns	14.683 ns	0.805 ns	1.01	0.01	-	NA
Kernel_Cjk	4	73.24 ns	11.233 ns	0.616 ns	0.99	0.01	-	NA
Dp	6	104.94 ns	3.524 ns	0.193 ns	1.00	0.00	-	NA
Kernel	6	80.35 ns	18.086 ns	0.991 ns	0.77	0.01	-	NA
Dp_Cjk	6	106.39 ns	40.766 ns	2.235 ns	1.01	0.02	-	NA
Kernel_Cjk	6	128.13 ns	9.392 ns	0.515 ns	1.22	0.00	-	NA
Dp	8	161.96 ns	2.060 ns	0.113 ns	1.00	0.00	-	NA
Kernel	8	92.55 ns	6.032 ns	0.331 ns	0.57	0.00	-	NA
Dp_Cjk	8	165.76 ns	33.273 ns	1.824 ns	1.02	0.01	-	NA
Kernel_Cjk	8	186.71 ns	13.578 ns	0.744 ns	1.15	0.00	-	NA
Dp	10	203.99 ns	4.454 ns	0.244 ns	1.00	0.00	-	NA
Kernel	10	102.39 ns	6.954 ns	0.381 ns	0.50	0.00	-	NA
Dp_Cjk	10	197.79 ns	17.841 ns	0.978 ns	0.97	0.00	-	NA
Kernel_Cjk	10	149.54 ns	7.521 ns	0.412 ns	0.73	0.00	-	NA
Dp	12	246.03 ns	9.615 ns	0.527 ns	1.00	0.00	-	NA
Kernel	12	108.59 ns	5.341 ns	0.293 ns	0.44	0.00	-	NA
Dp_Cjk	12	245.14 ns	5.370 ns	0.294 ns	1.00	0.00	-	NA
Kernel_Cjk	12	162.57 ns	3.161 ns	0.173 ns	0.66	0.00	-	NA
Dp	16	416.34 ns	178.404 ns	9.779 ns	1.00	0.03	-	NA
Kernel	16	133.11 ns	21.444 ns	1.175 ns	0.32	0.01	-	NA
Dp_Cjk	16	413.08 ns	140.171 ns	7.683 ns	0.99	0.03	-	NA
Kernel_Cjk	16	184.98 ns	75.911 ns	4.161 ns	0.44	0.01	-	NA
Dp	24	844.79 ns	59.856 ns	3.281 ns	1.00	0.00	-	NA
Kernel	24	160.06 ns	0.755 ns	0.041 ns	0.19	0.00	-	NA
Dp_Cjk	24	859.38 ns	129.940 ns	7.122 ns	1.02	0.01	-	NA
Kernel_Cjk	24	224.51 ns	1.123 ns	0.062 ns	0.27	0.00	-	NA
Dp	32	1,450.81 ns	16.071 ns	0.881 ns	1.00	0.00	-	NA
Kernel	32	197.41 ns	13.472 ns	0.738 ns	0.14	0.00	-	NA
Dp_Cjk	32	1,455.54 ns	162.573 ns	8.911 ns	1.00	0.01	-	NA
Kernel_Cjk	32	266.72 ns	1.641 ns	0.090 ns	0.18	0.00	-	NA
Dp	48	3,384.41 ns	2,289.327 ns	125.486 ns	1.00	0.04	-	NA
Kernel	48	268.96 ns	13.878 ns	0.761 ns	0.08	0.00	-	NA
Dp_Cjk	48	3,363.11 ns	870.020 ns	47.689 ns	0.99	0.03	-	NA
Kernel_Cjk	48	359.21 ns	32.588 ns	1.786 ns	0.11	0.00	-	NA
Dp	64	6,332.46 ns	2,537.791 ns	139.105 ns	1.00	0.03	-	NA
Kernel	64	343.32 ns	21.591 ns	1.183 ns	0.05	0.00	-	NA
Dp_Cjk	64	6,916.61 ns	5,559.145 ns	304.715 ns	1.09	0.05	-	NA
Kernel_Cjk	64	440.57 ns	21.123 ns	1.158 ns	0.07	0.00	-	NA
Dp	96	14,164.04 ns	3,011.813 ns	165.088 ns	1.00	0.01	-	NA
Kernel	96	738.77 ns	144.012 ns	7.894 ns	0.05	0.00	-	NA
Dp_Cjk	96	14,334.47 ns	10,473.812 ns	574.105 ns	1.01	0.04	-	NA
Kernel_Cjk	96	1,547.29 ns	669.977 ns	36.724 ns	0.11	0.00	-	NA

Lodestar.Text.Benchmarks.OsaBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_Utf16	8	39.77 ns	0.872 ns	0.048 ns	1.00	0.00	-	NA
Distance_CodePoint	8	125.21 ns	0.509 ns	0.028 ns	3.15	0.00	-	NA
Distance_Utf16	32	144.36 ns	1.472 ns	0.081 ns	1.00	0.00	-	NA
Distance_CodePoint	32	1,441.91 ns	842.781 ns	46.196 ns	9.99	0.28	-	NA
Distance_Utf16	64	494.38 ns	40.325 ns	2.210 ns	1.00	0.01	-	NA
Distance_CodePoint	64	9,046.63 ns	4,049.234 ns	221.952 ns	18.30	0.40	-	NA
Distance_Utf16	128	48,950.78 ns	16,947.603 ns	928.955 ns	1.00	0.02	-	NA
Distance_CodePoint	128	46,853.36 ns	5,585.059 ns	306.136 ns	0.96	0.02	-	NA

Lodestar.Text.Benchmarks.PartialFitBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	RowCount	BatchCount	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
WholeFit	10000	10	205.23 μs	8.574 μs	0.470 μs	1.00	0.00	-	800 B	1.00
BatchedFit	10000	10	209.01 μs	25.709 μs	1.409 μs	1.02	0.01	0.2441	7784 B	9.73
SparseFit	10000	10	59.17 μs	7.806 μs	0.428 μs	0.29	0.00	-	800 B	1.00
DenseFitOfTheSameData	10000	10	199.01 μs	48.503 μs	2.659 μs	0.97	0.01	-	800 B	1.00
WholeFit	10000	100	205.40 μs	48.544 μs	2.661 μs	1.00	0.02	-	800 B	1.00
BatchedFit	10000	100	232.62 μs	33.068 μs	1.813 μs	1.13	0.01	4.6387	77624 B	97.03
SparseFit	10000	100	58.91 μs	4.317 μs	0.237 μs	0.29	0.00	-	800 B	1.00
DenseFitOfTheSameData	10000	100	200.13 μs	71.807 μs	3.936 μs	0.97	0.02	-	800 B	1.00
WholeFit	100000	10	2,075.13 μs	517.371 μs	28.359 μs	1.00	0.02	-	803 B	1.00
BatchedFit	100000	10	2,035.15 μs	104.863 μs	5.748 μs	0.98	0.01	-	7787 B	9.70
SparseFit	100000	10	1,081.17 μs	92.501 μs	5.070 μs	0.52	0.01	-	800 B	1.00
DenseFitOfTheSameData	100000	10	1,970.18 μs	46.589 μs	2.554 μs	0.95	0.01	-	803 B	1.00
WholeFit	100000	100	1,969.68 μs	17.134 μs	0.939 μs	1.00	0.00	-	803 B	1.00
BatchedFit	100000	100	2,077.74 μs	140.101 μs	7.679 μs	1.05	0.00	3.9063	77627 B	96.67
SparseFit	100000	100	1,083.57 μs	75.504 μs	4.139 μs	0.55	0.00	-	801 B	1.00
DenseFitOfTheSameData	100000	100	1,988.47 μs	421.416 μs	23.099 μs	1.01	0.01	-	803 B	1.00

Lodestar.Text.Benchmarks.PartialRatioLongNeedleBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	NeedleLength	Mean	Error	StdDev	Allocated
Embedded	65	10.623 μs	0.9313 μs	0.0510 μs	-
EqualLength	65	2.992 μs	1.2071 μs	0.0662 μs	-
Embedded	128	40.172 μs	1.6329 μs	0.0895 μs	-
EqualLength	128	5.422 μs	0.4074 μs	0.0223 μs	-
Embedded	512	1,326.770 μs	299.6222 μs	16.4233 μs	-
EqualLength	512	43.254 μs	4.2797 μs	0.2346 μs	-

Lodestar.Text.Benchmarks.PartitionValidityBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
DaviesBouldinScore	1000000	11.06 ms	3.345 ms	0.183 ms	140.6250	140.6250	140.6250	3.82 MB
CalinskiHarabaszScore	1000000	16.59 ms	1.521 ms	0.083 ms	140.6250	140.6250	140.6250	3.82 MB

Lodestar.Text.Benchmarks.PersistenceBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
VocabTxt	3.996 ms	1.8464 ms	0.1012 ms	109.3750	101.5625	31.2500	3711.62 KB
TokenizerJsonWordPiece	11.547 ms	3.3845 ms	0.1855 ms	187.5000	171.8750	46.8750	5852.41 KB
TokenizerJsonUnigram	12.687 ms	0.2482 ms	0.0136 ms	78.1250	62.5000	15.6250	4748.83 KB
SpieceModel	5.500 ms	1.3627 ms	0.0747 ms	101.5625	93.7500	31.2500	4292.3 KB
TfidfSave	1.619 ms	0.9125 ms	0.0500 ms	21.4844	15.6250	15.6250	2137.05 KB
TfidfLoad	4.412 ms	0.1601 ms	0.0088 ms	85.9375	78.1250	23.4375	2921.38 KB
EmbeddingIndexSave	5.284 ms	0.6521 ms	0.0357 ms	257.8125	257.8125	257.8125	20349.96 KB
EmbeddingIndexLoad	8.387 ms	3.1169 ms	0.1708 ms	179.6875	148.4375	117.1875	16094.38 KB
EmbeddingIndexSaveFile	49.559 ms	7.7984 ms	0.4275 ms	-	-	-	323.43 KB
EmbeddingIndexLoadGzip	73.815 ms	5.1199 ms	0.2806 ms	-	-	-	16095.25 KB

Lodestar.Text.Benchmarks.PrecompiledNormalizerBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Allocated
NormalizeDocuments	13.57 ms	0.730 ms	0.040 ms	171.8750	2.75 MB
EncodeDocuments	61.33 ms	3.889 ms	0.213 ms	444.4444	8.19 MB

Lodestar.Text.Benchmarks.PrincipalComponentVarianceBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_ExplainedVariance	100x200	9,202.73 μs	1,460.147 μs	80.036 μs	1.00	0.01	46.8750	46.8750	46.8750	318.35 KB	1.00
NumFlat_Pca	100x200	15,299.65 μs	2,724.888 μs	149.360 μs	1.66	0.02	187.5000	187.5000	187.5000	628.54 KB	1.97
Lodestar_ExplainedVariance	2000x10	133.22 μs	3.642 μs	0.200 μs	1.00	0.00	-	-	-	2.45 KB	1.00
NumFlat_Pca	2000x10	180.83 μs	12.336 μs	0.676 μs	1.36	0.00	-	-	-	1.94 KB	0.79
Lodestar_ExplainedVariance	2000x50	2,808.86 μs	324.605 μs	17.793 μs	1.00	0.01	-	-	-	42.14 KB	1.00
NumFlat_Pca	2000x50	2,587.09 μs	182.817 μs	10.021 μs	0.92	0.01	-	-	-	40.07 KB	0.95
Lodestar_ExplainedVariance	200x10	22.80 μs	1.599 μs	0.088 μs	1.00	0.00	0.1221	-	-	2.45 KB	1.00
NumFlat_Pca	200x10	27.40 μs	2.523 μs	0.138 μs	1.20	0.01	0.0916	-	-	1.94 KB	0.79

Lodestar.Text.Benchmarks.ProcessExtractBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Limit	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
Extract	1	6.862 ms	0.5160 ms	0.0283 ms	1.00	118 B	1.00
ExtractOne	1	6.738 ms	0.1715 ms	0.0094 ms	0.98	-	0.00
Extract	5	6.850 ms	0.2531 ms	0.0139 ms	1.00	214 B	1.00
ExtractOne	5	6.742 ms	0.1639 ms	0.0090 ms	0.98	-	0.00

Lodestar.Text.Benchmarks.QgramBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Q	Element	Mean	Error	StdDev	Gen0	Allocated
JaccardSimilarity	1	Utf16Unit	4.792 μs	0.3528 μs	0.0193 μs	0.0076	192 B
JaccardSimilarity	1	CodePoint	3.949 μs	0.4233 μs	0.0232 μs	0.0076	192 B
JaccardSimilarity	3	Utf16Unit	5.695 μs	0.4756 μs	0.0261 μs	0.0076	192 B
JaccardSimilarity	3	CodePoint	5.072 μs	0.5430 μs	0.0298 μs	0.0076	192 B

Lodestar.Text.Benchmarks.RankingMetricsBenchmarks-report-github

As of 2026-11-02, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Rows	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
NdcgTieAveraged	100000	37.211 ms	1.6227 ms	0.0889 ms	-	-	-	800349 B
NdcgIgnoringTies	100000	35.029 ms	2.0170 ms	0.1106 ms	-	-	-	800345 B
DcgTieAveraged	100000	21.868 ms	3.4401 ms	0.1886 ms	-	-	-	800319 B
ReciprocalRankScore	100000	17.850 ms	0.8425 ms	0.0462 ms	-	-	-	-
CoverageErrorScore	100000	5.535 ms	0.1505 ms	0.0082 ms	15.6250	15.6250	15.6250	800040 B
LabelRankingAveragePrecisionScore	100000	53.414 ms	1.0417 ms	0.0571 ms	-	-	-	-

Lodestar.Text.Benchmarks.RatcliffObershelpBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Length	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Similarity_Containment	64	372.9 ns	18.62 ns	1.02 ns	1.00	0.00	-	NA
Similarity_NearDuplicate	64	5,366.8 ns	403.72 ns	22.13 ns	14.39	0.06	-	NA
Similarity_Containment	512	22,792.1 ns	4,594.02 ns	251.81 ns	1.00	0.01	-	NA
Similarity_NearDuplicate	512	833,518.5 ns	154,621.86 ns	8,475.35 ns	36.57	0.47	-	NA

Lodestar.Text.Benchmarks.RegressionMetricsBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Mse	100000	45.71 μs	7.276 μs	0.399 μs	-	-	-	-
Mae	100000	45.21 μs	15.112 μs	0.828 μs	-	-	-	-
R2Score	100000	110.23 μs	3.509 μs	0.192 μs	-	-	-	64 B
MedianAe	100000	546.48 μs	186.976 μs	10.249 μs	199.2188	199.2188	199.2188	800186 B
Mse	1000000	464.91 μs	98.268 μs	5.386 μs	-	-	-	-
Mae	1000000	451.61 μs	44.748 μs	2.453 μs	-	-	-	-
R2Score	1000000	1,110.18 μs	32.298 μs	1.770 μs	-	-	-	65 B
MedianAe	1000000	6,912.72 μs	513.648 μs	28.155 μs	320.3125	320.3125	320.3125	8000269 B

Lodestar.Text.Benchmarks.RobustScalerSparseBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Shape	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	2000x2000x20	26.95 ms	1.550 ms	0.085 ms	93.7500	93.7500	93.7500	367.51 KB
Fit	500x20000x5	44.21 ms	0.229 ms	0.013 ms	-	-	-	414.36 KB

Lodestar.Text.Benchmarks.ScalerIncumbentBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]     : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  Job-TSGYHD : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

LaunchCount=1

Method	Job	InvocationCount	IterationCount	RunStrategy	UnrollFactor	WarmupCount	RowCount	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_MinMax	Job-TSGYHD	1	20	Throughput	1	5	1000	110.221 μs	3.9999 μs	4.1076 μs	1.00	0.05	-	-	-	81632 B	1.00
Lodestar_MaxAbs	Job-TSGYHD	1	20	Throughput	1	5	1000	119.479 μs	19.4081 μs	21.5721 μs	1.09	0.20	-	-	-	81096 B	0.99
Lodestar_Robust	Job-TSGYHD	1	20	Throughput	1	5	1000	2,296.173 μs	26.4201 μs	30.4255 μs	20.86	0.82	-	-	-	161352 B	1.98
Lodestar_Standard	Job-TSGYHD	1	20	Throughput	1	5	1000	123.460 μs	23.6193 μs	27.2000 μs	1.12	0.24	-	-	-	81560 B	1.00
Lodestar_Standard_SparseFit	Job-TSGYHD	1	20	Throughput	1	5	1000	77.585 μs	12.5319 μs	13.4090 μs	0.70	0.12	-	-	-	-	0.00
Lodestar_MaxAbs_SparseFit	Job-TSGYHD	1	20	Throughput	1	5	1000	67.220 μs	12.6848 μs	13.5726 μs	0.61	0.12	-	-	-	-	0.00
Lodestar_MaxAbs_SparseClip	Job-TSGYHD	1	20	Throughput	1	5	1000	104.644 μs	15.9574 μs	15.6723 μs	0.95	0.14	-	-	-	27856 B	0.34
MlNet_NormalizeMinMax_Fit	Job-TSGYHD	1	20	Throughput	1	5	1000	440.296 μs	16.8253 μs	18.7013 μs	4.00	0.22	-	-	-	9528 B	0.12
MlNet_NormalizeMinMax_Read	Job-TSGYHD	1	20	Throughput	1	5	1000	2,270.320 μs	87.4649 μs	97.2170 μs	20.63	1.15	-	-	-	297008 B	3.64
MlNet_NormalizeRobustScaling_Read	Job-TSGYHD	1	20	Throughput	1	5	1000	6,477.092 μs	128.5983 μs	132.0610 μs	58.84	2.48	-	-	-	461192 B	5.65
Lodestar_MinMax	ShortRun	Default	3	Default	16	3	1000	48.724 μs	6.8228 μs	0.3740 μs	1.00	0.01	4.7607	-	-	80896 B	1.000
Lodestar_MaxAbs	ShortRun	Default	3	Default	16	3	1000	41.184 μs	6.2599 μs	0.3431 μs	0.85	0.01	4.7607	-	-	80336 B	0.993
Lodestar_Robust	ShortRun	Default	3	Default	16	3	1000	465.402 μs	15.4942 μs	0.8493 μs	9.55	0.07	9.2773	-	-	160616 B	1.985
Lodestar_Standard	ShortRun	Default	3	Default	16	3	1000	46.167 μs	1.5703 μs	0.0861 μs	0.95	0.01	4.7607	-	-	80824 B	0.999
Lodestar_Standard_SparseFit	ShortRun	Default	3	Default	16	3	1000	10.565 μs	0.7846 μs	0.0430 μs	0.22	0.00	0.0458	-	-	824 B	0.010
Lodestar_MaxAbs_SparseFit	ShortRun	Default	3	Default	16	3	1000	7.681 μs	0.7368 μs	0.0404 μs	0.16	0.00	0.0153	-	-	376 B	0.005
Lodestar_MaxAbs_SparseClip	ShortRun	Default	3	Default	16	3	1000	11.925 μs	2.1837 μs	0.1197 μs	0.24	0.00	1.6174	0.1068	-	27120 B	0.335
MlNet_NormalizeMinMax_Fit	ShortRun	Default	3	Default	16	3	1000	1,564.062 μs	4,361.1427 μs	239.0489 μs	32.10	4.25	83.4961	7.3242	-	1398987 B	17.294
MlNet_NormalizeMinMax_Read	ShortRun	Default	3	Default	16	3	1000	1,062.521 μs	1,380.9001 μs	75.6918 μs	21.81	1.35	39.0625	1.9531	-	652764 B	8.069
MlNet_NormalizeRobustScaling_Read	ShortRun	Default	3	Default	16	3	1000	1,644.308 μs	2,783.0052 μs	152.5459 μs	33.75	2.72	35.1563	-	-	625462 B	7.732
Lodestar_MinMax	Job-TSGYHD	1	20	Throughput	1	5	20000	1,545.853 μs	14.5609 μs	16.7683 μs	1.00	0.01	-	-	-	1601632 B	1.000
Lodestar_MaxAbs	Job-TSGYHD	1	20	Throughput	1	5	20000	1,451.713 μs	11.0741 μs	12.3089 μs	0.94	0.01	-	-	-	1601096 B	1.000
Lodestar_Robust	Job-TSGYHD	1	20	Throughput	1	5	20000	15,236.034 μs	77.9333 μs	83.3877 μs	9.86	0.12	-	-	-	3201352 B	1.999
Lodestar_Standard	Job-TSGYHD	1	20	Throughput	1	5	20000	1,541.842 μs	17.3956 μs	20.0329 μs	1.00	0.02	-	-	-	1601560 B	1.000
Lodestar_Standard_SparseFit	Job-TSGYHD	1	20	Throughput	1	5	20000	417.040 μs	7.2946 μs	8.1080 μs	0.27	0.01	-	-	-	-	0.000
Lodestar_MaxAbs_SparseFit	Job-TSGYHD	1	20	Throughput	1	5	20000	310.750 μs	2.1239 μs	1.9867 μs	0.20	0.00	-	-	-	-	0.000
Lodestar_MaxAbs_SparseClip	Job-TSGYHD	1	20	Throughput	1	5	20000	428.206 μs	11.9083 μs	13.2361 μs	0.28	0.01	-	-	-	559752 B	0.349
MlNet_NormalizeMinMax_Fit	Job-TSGYHD	1	20	Throughput	1	5	20000	6,764.834 μs	41.6950 μs	46.3438 μs	4.38	0.05	-	-	-	9528 B	0.006
MlNet_NormalizeMinMax_Read	Job-TSGYHD	1	20	Throughput	1	5	20000	17,660.220 μs	8,421.3042 μs	9,697.9864 μs	11.43	6.12	-	-	-	530424 B	0.331
MlNet_NormalizeRobustScaling_Read	Job-TSGYHD	1	20	Throughput	1	5	20000	19,310.536 μs	194.5762 μs	208.1943 μs	12.49	0.19	-	-	-	4278520 B	2.671
Lodestar_MinMax	ShortRun	Default	3	Default	16	3	20000	1,386.666 μs	1,663.9432 μs	91.2063 μs	1.00	0.08	330.0781	330.0781	330.0781	1601938 B	1.000
Lodestar_MaxAbs	ShortRun	Default	3	Default	16	3	20000	1,433.928 μs	729.0498 μs	39.9617 μs	1.04	0.06	330.0781	330.0781	330.0781	1600551 B	0.999
Lodestar_Robust	ShortRun	Default	3	Default	16	3	20000	14,479.966 μs	98.9343 μs	5.4229 μs	10.47	0.58	875.0000	875.0000	875.0000	3201194 B	1.998
Lodestar_Standard	ShortRun	Default	3	Default	16	3	20000	1,465.670 μs	1,467.8257 μs	80.4565 μs	1.06	0.08	330.0781	330.0781	330.0781	1601039 B	0.999
Lodestar_Standard_SparseFit	ShortRun	Default	3	Default	16	3	20000	298.243 μs	79.2462 μs	4.3438 μs	0.22	0.01	-	-	-	824 B	0.001
Lodestar_MaxAbs_SparseFit	ShortRun	Default	3	Default	16	3	20000	244.214 μs	12.8615 μs	0.7050 μs	0.18	0.01	-	-	-	376 B	0.000
Lodestar_MaxAbs_SparseClip	ShortRun	Default	3	Default	16	3	20000	402.342 μs	35.1962 μs	1.9292 μs	0.29	0.02	128.9063	124.0234	124.0234	560068 B	0.350
MlNet_NormalizeMinMax_Fit	ShortRun	Default	3	Default	16	3	20000	1,633.497 μs	2,378.5702 μs	130.3775 μs	1.18	0.10	41.0156	6.8359	-	702655 B	0.439
MlNet_NormalizeMinMax_Read	ShortRun	Default	3	Default	16	3	20000	4,413.796 μs	1,620.0944 μs	88.8028 μs	3.19	0.19	31.2500	7.8125	-	594265 B	0.371
MlNet_NormalizeRobustScaling_Read	ShortRun	Default	3	Default	16	3	20000	19,463.651 μs	600.5423 μs	32.9177 μs	14.08	0.78	250.0000	125.0000	-	4302324 B	2.686

Lodestar.Text.Benchmarks.SentencePieceBpeLineageBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Model	Mean	Error	StdDev	Gen0	Allocated
EncodeDocuments	Llama2	137.9 ms	5.28 ms	0.29 ms	2250.0000	37.05 MB
EncodeDocuments	Mistral	134.1 ms	20.37 ms	1.12 ms	2250.0000	37.03 MB

Lodestar.Text.Benchmarks.SilhouetteBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Samples	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
PerSample	2000	27.30 ms	0.536 ms	0.029 ms	31.2500	31.2500	31.2500	148.75 KB
PerSample	5000	193.75 ms	3.188 ms	0.175 ms	-	-	-	371.6 KB

Lodestar.Text.Benchmarks.SimilaritySketchBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Permutations	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
ExactPairwise	500	64	41.758 ms	1.0502 ms	0.0576 ms	1.00	0.00	416.6667	-	-	8152.4 KB	1.00
SketchThenVerify	500	64	8.632 ms	1.1539 ms	0.0632 ms	0.21	0.00	171.8750	156.2500	78.1250	2533.15 KB	0.31
SignaturesOnly	500	64	6.817 ms	0.2308 ms	0.0127 ms	0.16	0.00	15.6250	-	-	304.73 KB	0.04
AffineSignaturesOnly	500	64	6.496 ms	0.5918 ms	0.0324 ms	0.16	0.00	15.6250	-	-	305.28 KB	0.04
FingerprintsOnly	500	64	8.168 ms	0.6065 ms	0.0332 ms	0.20	0.00	15.6250	-	-	429.7 KB	0.05
ExactPairwise	500	128	43.496 ms	4.6141 ms	0.2529 ms	1.00	0.01	416.6667	-	-	8152.4 KB	1.00
SketchThenVerify	500	128	11.287 ms	0.6029 ms	0.0330 ms	0.26	0.00	328.1250	312.5000	218.7500	4578.87 KB	0.56
SignaturesOnly	500	128	8.222 ms	0.9068 ms	0.0497 ms	0.19	0.00	15.6250	-	-	429.73 KB	0.05
AffineSignaturesOnly	500	128	5.405 ms	0.8045 ms	0.0441 ms	0.12	0.00	23.4375	-	-	430.78 KB	0.05
FingerprintsOnly	500	128	8.242 ms	2.8085 ms	0.1539 ms	0.19	0.00	15.6250	-	-	429.7 KB	0.05
ExactPairwise	2000	64	759.450 ms	145.6830 ms	7.9854 ms	1.00	0.01	7000.0000	-	-	126360.09 KB	1.000
SketchThenVerify	2000	64	36.862 ms	0.6983 ms	0.0383 ms	0.05	0.00	928.5714	928.5714	500.0000	10335.83 KB	0.082
SignaturesOnly	2000	64	31.703 ms	3.0283 ms	0.1660 ms	0.04	0.00	62.5000	-	-	1218.83 KB	0.010
AffineSignaturesOnly	2000	64	21.370 ms	1.5878 ms	0.0870 ms	0.03	0.00	62.5000	-	-	1219.36 KB	0.010
FingerprintsOnly	2000	64	32.528 ms	1.2248 ms	0.0671 ms	0.04	0.00	62.5000	-	-	1718.75 KB	0.014
ExactPairwise	2000	128	771.191 ms	261.0651 ms	14.3099 ms	1.00	0.02	7000.0000	-	-	126360.09 KB	1.00
SketchThenVerify	2000	128	58.362 ms	17.0622 ms	0.9352 ms	0.08	0.00	1555.5556	1555.5556	888.8889	18665.05 KB	0.15
SignaturesOnly	2000	128	37.274 ms	0.2958 ms	0.0162 ms	0.05	0.00	71.4286	-	-	1718.84 KB	0.01
AffineSignaturesOnly	2000	128	21.806 ms	0.3299 ms	0.0181 ms	0.03	0.00	93.7500	-	-	1719.86 KB	0.01
FingerprintsOnly	2000	128	37.374 ms	0.3459 ms	0.0190 ms	0.05	0.00	71.4286	-	-	1718.8 KB	0.01

Lodestar.Text.Benchmarks.SmallArtifactBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
CountSaveAsync	1,196.8 ns	149.1 ns	8.17 ns	0.3262	0.0057	-	5464 B
TfidfSaveAsync	1,445.9 ns	479.7 ns	26.29 ns	0.3338	-	-	5592 B
HashingSaveAsync	1,228.8 ns	478.8 ns	26.24 ns	0.3281	0.0057	-	5496 B
IndexSaveAsync	472.3 ns	109.4 ns	5.99 ns	0.0386	-	-	648 B
CountLoadSeekable1400K	12,177,124.7 ns	1,549,748.5 ns	84,946.95 ns	890.6250	859.3750	625.0000	12227088 B

Lodestar.Text.Benchmarks.SplitterIncumbentBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	SampleCount	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_KFold	10000	79.000 μs	31.907 μs	1.7489 μs	1.00	0.03	16.7236	4.7607	-	274.17 KB	1.00
Lodestar_StratifiedKFold	10000	189.465 μs	42.886 μs	2.3507 μs	2.40	0.05	19.0430	6.8359	-	313.77 KB	1.14
Lodestar_TrainTest	10000	4.795 μs	1.183 μs	0.0648 μs	0.06	0.00	2.3956	0.2594	-	39.19 KB	0.14
MlNet_CrossValidationSplit	10000	332.459 μs	34.246 μs	1.8771 μs	4.21	0.08	2.4414	0.9766	-	40.55 KB	0.15
MlNet_CrossValidationSplit_Read	10000	6,845.573 μs	771.034 μs	42.2630 μs	86.68	1.71	7.8125	-	-	149.06 KB	0.54
MlNet_TrainTestSplit	10000	64.248 μs	7.287 μs	0.3994 μs	0.81	0.02	0.7324	0.2441	-	12.55 KB	0.05
MlNet_TrainTestSplit_Read	10000	1,154.760 μs	569.908 μs	31.2386 μs	14.62	0.44	1.9531	-	-	34.26 KB	0.12
Lodestar_KFold	100000	1,104.476 μs	271.748 μs	14.8954 μs	1.00	0.02	458.9844	445.3125	435.5469	2738.56 KB	1.000
Lodestar_StratifiedKFold	100000	2,417.115 μs	619.176 μs	33.9391 μs	2.19	0.04	492.1875	476.5625	468.7500	3130.08 KB	1.143
Lodestar_TrainTest	100000	82.197 μs	63.733 μs	3.4934 μs	0.07	0.00	52.4902	52.4902	52.4902	391.17 KB	0.143
MlNet_CrossValidationSplit	100000	281.979 μs	184.385 μs	10.1068 μs	0.26	0.01	1.9531	-	-	40.55 KB	0.015
MlNet_CrossValidationSplit_Read	100000	44,128.260 μs	4,712.400 μs	258.3026 μs	39.96	0.51	-	-	-	148.79 KB	0.054
MlNet_TrainTestSplit	100000	64.410 μs	3.142 μs	0.1722 μs	0.06	0.00	0.7324	0.2441	-	12.55 KB	0.005
MlNet_TrainTestSplit_Read	100000	7,674.297 μs	598.047 μs	32.7810 μs	6.95	0.09	-	-	-	34.19 KB	0.012

Lodestar.Text.Benchmarks.StopWordBenchmarks-report-github

As of 2026-10-30, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Mean	Error	StdDev	Ratio	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Count	200	1.872 ms	0.2206 ms	0.0121 ms	1.00	115.2344	115.2344	115.2344	869.91 KB	1.00
CountWithStopWords	200	1.387 ms	0.0924 ms	0.0051 ms	0.74	23.4375	7.8125	-	407.6 KB	0.47
Hashing	200	1.994 ms	0.0767 ms	0.0042 ms	1.07	70.3125	70.3125	70.3125	554.26 KB	0.64
HashingWithStopWords	200	1.707 ms	0.1451 ms	0.0080 ms	0.91	15.6250	3.9063	-	273.65 KB	0.31
Count	1000	8.849 ms	0.2746 ms	0.0151 ms	1.00	859.3750	859.3750	859.3750	3436.53 KB	1.00
CountWithStopWords	1000	7.075 ms	0.2018 ms	0.0111 ms	0.80	382.8125	382.8125	382.8125	1660.72 KB	0.48
Hashing	1000	9.556 ms	0.1343 ms	0.0074 ms	1.08	531.2500	515.6250	515.6250	2349.23 KB	0.68
HashingWithStopWords	1000	8.688 ms	0.2021 ms	0.0111 ms	0.98	265.6250	265.6250	265.6250	1139.46 KB	0.33

Lodestar.Text.Benchmarks.TextRankBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Words	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Extract	2000	650.3 μs	128.1 μs	7.02 μs	30.2734	10.7422	-	499.33 KB
Extract	8000	2,789.6 μs	138.8 μs	7.61 μs	121.0938	78.1250	39.0625	1965.54 KB
Extract	32000	12,144.6 μs	2,254.5 μs	123.58 μs	859.3750	859.3750	734.3750	7946.83 KB

Lodestar.Text.Benchmarks.TokenPatternBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Pattern	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
Fit	[^ ]+	28,303,004.5 ns	524,638.13 ns	28,757.19 ns	62.5000	31.2500	-	1554.58 KB
Construct	[^ ]+	1,494.5 ns	59.73 ns	3.27 ns	0.3319	0.0019	-	5.43 KB
ConstructAndFitOneDocument	[^ ]+	2,495.1 ns	268.26 ns	14.70 ns	0.4539	0.0038	-	7.43 KB
Fit	[a-z]+	24,073,506.6 ns	648,576.53 ns	35,550.67 ns	-	-	-	7.2 KB
Construct	[a-z]+	1,442.0 ns	403.00 ns	22.09 ns	0.3300	0.0019	-	5.4 KB
ConstructAndFitOneDocument	[a-z]+	2,173.4 ns	278.54 ns	15.27 ns	0.4387	0.0038	-	7.18 KB
Fit	[A-Za-z]\w+	32,385,570.6 ns	750,415.47 ns	41,132.80 ns	62.5000	-	-	1555.98 KB
Construct	[A-Za-z]\w+	2,687.4 ns	524.72 ns	28.76 ns	0.4158	0.0038	-	6.81 KB
ConstructAndFitOneDocument	[A-Za-z]\w+	3,624.1 ns	1,070.47 ns	58.68 ns	0.5188	-	-	8.81 KB
Fit	\b\w\w+\b	16,515,674.2 ns	775,166.63 ns	42,489.50 ns	125.0000	93.7500	62.5000	9743.56 KB
Construct	\b\w\w+\b	831.7 ns	57.52 ns	3.15 ns	0.1135	-	-	1.87 KB
ConstructAndFitOneDocument	\b\w\w+\b	1,536.9 ns	73.77 ns	4.04 ns	0.2346	-	-	3.84 KB
Fit	\S+	29,176,210.6 ns	2,321,165.15 ns	127,230.89 ns	62.5000	31.2500	-	1554.45 KB
Construct	\S+	1,639.6 ns	86.08 ns	4.72 ns	0.3223	0.0019	-	5.29 KB
ConstructAndFitOneDocument	\S+	2,581.1 ns	305.30 ns	16.73 ns	0.4425	0.0038	-	7.29 KB

Lodestar.Text.Benchmarks.TokenizerIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Model	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Lodestar	WordPiece	20.82 ms	0.844 ms	0.046 ms	1.00	0.00	531.2500	8.71 MB	1.00
MlTokenizers	WordPiece	53.94 ms	6.963 ms	0.382 ms	2.59	0.02	200.0000	3.55 MB	0.41
Lodestar	SentencePiece	63.40 ms	5.916 ms	0.324 ms	1.00	0.01	500.0000	7.98 MB	1.00
MlTokenizers	SentencePiece	68.55 ms	33.633 ms	1.844 ms	1.08	0.03	250.0000	5.65 MB	0.71
Lodestar	ByteLevelBpe	86.41 ms	29.799 ms	1.633 ms	1.00	0.02	1666.6667	28.47 MB	1.00
MlTokenizers	ByteLevelBpe	255.04 ms	31.931 ms	1.750 ms	2.95	0.05	3500.0000	59.08 MB	2.08

Lodestar.Text.Benchmarks.TopKAccuracyBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Classes	Mean	Error	StdDev	Allocated
TopTwo	10	11.68 ms	5.008 ms	0.275 ms	-
TopTwo	100	87.64 ms	7.066 ms	0.387 ms	-

Lodestar.Text.Benchmarks.TransformerIncumbentBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]     : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  Job-UDACYX : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

LaunchCount=1

Method	Job	InvocationCount	IterationCount	RunStrategy	UnrollFactor	WarmupCount	RowCount	Mean	Error	StdDev	Median	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar_Normalize	Job-UDACYX	1	20	Throughput	1	5	1000	47.42 μs	5.275 μs	5.644 μs	46.06 μs	1.01	0.16	-	-	-	31.99 KB	1.00
Lodestar_Polynomial	Job-UDACYX	1	20	Throughput	1	5	1000	82.03 μs	4.526 μs	5.031 μs	80.94 μs	1.75	0.21	-	-	-	118.16 KB	3.69
Lodestar_Discretize	Job-UDACYX	1	20	Throughput	1	5	1000	243.73 μs	20.008 μs	23.042 μs	245.32 μs	5.20	0.72	-	-	-	156.99 KB	4.91
Lodestar_Discretize_64Bins	Job-UDACYX	1	20	Throughput	1	5	1000	273.53 μs	13.996 μs	16.117 μs	267.00 μs	5.84	0.68	-	-	-	31.99 KB	1.00
Lodestar_Discretize_Fit	Job-UDACYX	1	20	Throughput	1	5	1000	999.18 μs	18.847 μs	20.948 μs	993.76 μs	21.32	2.21	-	-	-	33.98 KB	1.06
Lodestar_Quantile	Job-UDACYX	1	20	Throughput	1	5	1000	1,632.43 μs	20.220 μs	22.475 μs	1,630.44 μs	34.83	3.58	-	-	-	31.99 KB	1.00
Lodestar_Quantile_Fit	Job-UDACYX	1	20	Throughput	1	5	1000	1,087.16 μs	13.226 μs	13.582 μs	1,087.68 μs	23.20	2.38	-	-	-	110.84 KB	3.46
Lodestar_Power	Job-UDACYX	1	20	Throughput	1	5	1000	219.39 μs	12.833 μs	14.779 μs	213.02 μs	4.68	0.57	-	-	-	31.99 KB	1.00
Lodestar_Power_Fit	Job-UDACYX	1	20	Throughput	1	5	1000	1,898.50 μs	19.069 μs	21.196 μs	1,896.51 μs	40.51	4.15	-	-	-	135.77 KB	4.24
Lodestar_KnnImpute	Job-UDACYX	1	20	Throughput	1	5	1000	14,245.99 μs	5,041.560 μs	5,805.868 μs	10,179.52 μs	303.96	125.41	-	-	-	103.87 KB	3.25
MlNet_NormalizeLpNorm_Read	Job-UDACYX	1	20	Throughput	1	5	1000	1,217.14 μs	94.421 μs	96.964 μs	1,240.46 μs	25.97	3.33	-	-	-	315.34 KB	9.86
MlNet_NormalizeBinning_Read	Job-UDACYX	1	20	Throughput	1	5	1000	3,099.85 μs	143.964 μs	154.040 μs	3,137.43 μs	66.14	7.46	-	-	-	401.91 KB	12.56
Lodestar_Normalize	ShortRun	Default	3	Default	16	3	1000	14.01 μs	0.504 μs	0.028 μs	14.02 μs	1.00	0.00	1.9073	-	-	31.27 KB	1.00
Lodestar_Polynomial	ShortRun	Default	3	Default	16	3	1000	83.50 μs	18.449 μs	1.011 μs	83.21 μs	5.96	0.06	36.9873	36.9873	36.9873	117.47 KB	3.76
Lodestar_Discretize	ShortRun	Default	3	Default	16	3	1000	137.57 μs	14.006 μs	0.768 μs	137.68 μs	9.82	0.05	49.8047	49.8047	49.8047	156.31 KB	5.00
Lodestar_Discretize_64Bins	ShortRun	Default	3	Default	16	3	1000	74.23 μs	1.980 μs	0.109 μs	74.21 μs	5.30	0.01	1.8311	-	-	31.27 KB	1.00
Lodestar_Discretize_Fit	ShortRun	Default	3	Default	16	3	1000	87.59 μs	200.948 μs	11.015 μs	93.40 μs	6.25	0.68	1.9531	-	-	33.26 KB	1.06
Lodestar_Quantile	ShortRun	Default	3	Default	16	3	1000	462.76 μs	45.894 μs	2.516 μs	463.29 μs	33.02	0.17	1.4648	-	-	31.27 KB	1.00
Lodestar_Quantile_Fit	ShortRun	Default	3	Default	16	3	1000	196.52 μs	60.490 μs	3.316 μs	195.25 μs	14.02	0.21	6.5918	0.9766	-	110.1 KB	3.52
Lodestar_Power	ShortRun	Default	3	Default	16	3	1000	140.50 μs	2.199 μs	0.121 μs	140.44 μs	10.03	0.02	1.7090	-	-	31.27 KB	1.00
Lodestar_Power_Fit	ShortRun	Default	3	Default	16	3	1000	721.87 μs	18.562 μs	1.017 μs	722.29 μs	51.51	0.11	7.8125	-	-	134.99 KB	4.32
Lodestar_KnnImpute	ShortRun	Default	3	Default	16	3	1000	8,722.16 μs	577.275 μs	31.642 μs	8,707.06 μs	622.41	2.23	-	-	-	103.16 KB	3.30
MlNet_NormalizeLpNorm_Read	ShortRun	Default	3	Default	16	3	1000	254.17 μs	49.114 μs	2.692 μs	253.29 μs	18.14	0.17	19.0430	18.5547	0.4883	297.69 KB	9.52
MlNet_NormalizeBinning_Read	ShortRun	Default	3	Default	16	3	1000	982.44 μs	1,175.145 μs	64.414 μs	971.23 μs	70.11	3.98	42.9688	5.8594	-	698.95 KB	22.35
Lodestar_Normalize	Job-UDACYX	1	20	Throughput	1	5	20000	479.50 μs	11.269 μs	12.977 μs	479.03 μs	1.00	0.04	-	-	-	625.74 KB	1.00
Lodestar_Polynomial	Job-UDACYX	1	20	Throughput	1	5	20000	1,655.58 μs	12.738 μs	13.629 μs	1,651.38 μs	3.46	0.09	-	-	-	2344.73 KB	3.75
Lodestar_Discretize	Job-UDACYX	1	20	Throughput	1	5	20000	3,575.06 μs	20.853 μs	24.015 μs	3,567.09 μs	7.46	0.20	-	-	-	3125.74 KB	5.00
Lodestar_Discretize_64Bins	Job-UDACYX	1	20	Throughput	1	5	20000	3,886.45 μs	962.966 μs	1,108.954 μs	4,465.50 μs	8.11	2.27	-	-	-	625.74 KB	1.00
Lodestar_Discretize_Fit	Job-UDACYX	1	20	Throughput	1	5	20000	11,669.42 μs	5,762.901 μs	6,636.565 μs	6,241.56 μs	24.35	13.54	-	-	-	627.73 KB	1.00
Lodestar_Quantile	Job-UDACYX	1	20	Throughput	1	5	20000	10,878.86 μs	36.820 μs	39.397 μs	10,872.37 μs	22.70	0.60	-	-	-	625.74 KB	1.00
Lodestar_Quantile_Fit	Job-UDACYX	1	20	Throughput	1	5	20000	9,642.81 μs	5,222.015 μs	6,013.680 μs	5,813.72 μs	20.12	12.26	-	-	-	704.59 KB	1.13
Lodestar_Power	Job-UDACYX	1	20	Throughput	1	5	20000	3,083.43 μs	21.758 μs	25.057 μs	3,083.37 μs	6.43	0.18	-	-	-	625.74 KB	1.00
Lodestar_Power_Fit	Job-UDACYX	1	20	Throughput	1	5	20000	14,680.35 μs	135.792 μs	150.932 μs	14,728.15 μs	30.64	0.86	-	-	-	2659.2 KB	4.25
Lodestar_KnnImpute	Job-UDACYX	1	20	Throughput	1	5	20000	9,696.00 μs	56.740 μs	58.267 μs	9,673.02 μs	20.24	0.54	-	-	-	105.37 KB	0.17
MlNet_NormalizeLpNorm_Read	Job-UDACYX	1	20	Throughput	1	5	20000	12,110.09 μs	1,773.907 μs	1,898.060 μs	13,118.57 μs	25.27	3.91	-	-	-	546.52 KB	0.87
MlNet_NormalizeBinning_Read	Job-UDACYX	1	20	Throughput	1	5	20000	29,371.60 μs	14,122.988 μs	16,264.054 μs	18,933.46 μs	61.30	33.17	-	-	-	2078.1 KB	3.32
Lodestar_Normalize	ShortRun	Default	3	Default	16	3	20000	680.90 μs	103.557 μs	5.676 μs	681.63 μs	1.00	0.01	199.2188	199.2188	199.2188	625.16 KB	1.00
Lodestar_Polynomial	ShortRun	Default	3	Default	16	3	20000	653.81 μs	380.520 μs	20.858 μs	655.12 μs	0.96	0.03	282.2266	282.2266	282.2266	2346.24 KB	3.75
Lodestar_Discretize	ShortRun	Default	3	Default	16	3	20000	3,167.27 μs	304.184 μs	16.673 μs	3,167.32 μs	4.65	0.04	996.0938	996.0938	996.0938	3125.68 KB	5.00
Lodestar_Discretize_64Bins	ShortRun	Default	3	Default	16	3	20000	1,894.95 μs	74.592 μs	4.089 μs	1,896.89 μs	2.78	0.02	199.2188	199.2188	199.2188	625.16 KB	1.00
Lodestar_Discretize_Fit	ShortRun	Default	3	Default	16	3	20000	5,742.40 μs	61.567 μs	3.375 μs	5,741.82 μs	8.43	0.06	195.3125	195.3125	195.3125	627.14 KB	1.00
Lodestar_Quantile	ShortRun	Default	3	Default	16	3	20000	10,680.29 μs	595.405 μs	32.636 μs	10,673.43 μs	15.69	0.12	187.5000	187.5000	187.5000	625.16 KB	1.00
Lodestar_Quantile_Fit	ShortRun	Default	3	Default	16	3	20000	5,889.44 μs	209.120 μs	11.463 μs	5,894.54 μs	8.65	0.06	195.3125	195.3125	195.3125	703.99 KB	1.13
Lodestar_Power	ShortRun	Default	3	Default	16	3	20000	3,208.37 μs	201.180 μs	11.027 μs	3,204.29 μs	4.71	0.04	199.2188	199.2188	199.2188	625.16 KB	1.00
Lodestar_Power_Fit	ShortRun	Default	3	Default	16	3	20000	13,748.99 μs	615.462 μs	33.736 μs	13,755.10 μs	20.19	0.15	750.0000	750.0000	750.0000	2660.98 KB	4.26
Lodestar_KnnImpute	ShortRun	Default	3	Default	16	3	20000	9,151.77 μs	381.146 μs	20.892 μs	9,141.18 μs	13.44	0.10	-	-	-	104.66 KB	0.17
MlNet_NormalizeLpNorm_Read	ShortRun	Default	3	Default	16	3	20000	2,288.57 μs	1,090.981 μs	59.800 μs	2,283.81 μs	3.36	0.08	27.3438	11.7188	-	454.68 KB	0.73
MlNet_NormalizeBinning_Read	ShortRun	Default	3	Default	16	3	20000	10,670.80 μs	633.926 μs	34.748 μs	10,659.19 μs	15.67	0.12	265.6250	218.7500	187.5000	2122.9 KB	3.40

Lodestar.Text.Benchmarks.VectorMathBenchmarks-report-github

As of 2026-10-30, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Dim	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
Dot	384	48.26 ns	5.224 ns	0.286 ns	1.00	-	NA
L2Norm	384	49.43 ns	0.881 ns	0.048 ns	1.02	-	NA
Dot	768	95.29 ns	3.078 ns	0.169 ns	1.00	-	NA
L2Norm	768	93.51 ns	0.973 ns	0.053 ns	0.98	-	NA
Dot	1024	124.09 ns	5.627 ns	0.308 ns	1.00	-	NA
L2Norm	1024	123.41 ns	1.620 ns	0.089 ns	0.99	-	NA

Lodestar.Text.Benchmarks.VectorStoreChurnBenchmarks-report-github

As of 2026-11-01, measured at commit 3e48e17c68bff96a3371172656990689433f479e.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]     : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  Job-AKIUDK : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2
  ShortRun   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX2

LaunchCount=1

Method	Job	IterationCount	RunStrategy	WarmupCount	Records	Mean	Error	StdDev	Gen0	Allocated
UpsertThenSearch	Job-AKIUDK	20	Throughput	5	1000	61.40 μs	0.054 μs	0.056 μs	0.1221	3.15 KB
UpsertThenHybrid	Job-AKIUDK	20	Throughput	5	1000	76.65 μs	0.183 μs	0.203 μs	1.2207	20.58 KB
Hybrid	Job-AKIUDK	20	Throughput	5	1000	70.69 μs	0.097 μs	0.103 μs	0.8545	15.52 KB
UpsertThenSearch	ShortRun	3	Default	3	1000	59.63 μs	2.083 μs	0.114 μs	0.1831	3.15 KB
UpsertThenHybrid	ShortRun	3	Default	3	1000	75.64 μs	3.017 μs	0.165 μs	1.2207	20.58 KB
Hybrid	ShortRun	3	Default	3	1000	71.50 μs	4.410 μs	0.242 μs	0.8545	15.52 KB
UpsertThenSearch	Job-AKIUDK	20	Throughput	5	10000	551.78 μs	1.342 μs	1.436 μs	-	3.15 KB
UpsertThenHybrid	Job-AKIUDK	20	Throughput	5	10000	894.21 μs	10.194 μs	11.330 μs	0.9766	26.16 KB
Hybrid	Job-AKIUDK	20	Throughput	5	10000	829.25 μs	2.793 μs	2.988 μs	0.9766	21.14 KB
UpsertThenSearch	ShortRun	3	Default	3	10000	581.02 μs	193.700 μs	10.617 μs	-	3.15 KB
UpsertThenHybrid	ShortRun	3	Default	3	10000	854.67 μs	59.998 μs	3.289 μs	0.9766	26.16 KB
Hybrid	ShortRun	3	Default	3	10000	866.59 μs	68.713 μs	3.766 μs	0.9766	21.14 KB
UpsertThenSearch	Job-AKIUDK	20	Throughput	5	100000	8,617.58 μs	58.498 μs	67.366 μs	-	3.16 KB
UpsertThenHybrid	Job-AKIUDK	20	Throughput	5	100000	12,611.72 μs	72.666 μs	83.682 μs	-	46 KB
Hybrid	Job-AKIUDK	20	Throughput	5	100000	12,675.67 μs	97.449 μs	112.222 μs	-	41.02 KB
UpsertThenSearch	ShortRun	3	Default	3	100000	8,822.61 μs	308.219 μs	16.895 μs	-	3.19 KB
UpsertThenHybrid	ShortRun	3	Default	3	100000	12,648.13 μs	1,842.959 μs	101.019 μs	-	46 KB
Hybrid	ShortRun	3	Default	3	100000	12,842.94 μs	1,514.929 μs	83.038 μs	-	41.02 KB

Lodestar.Text.Benchmarks.VectorizerBenchmarks-report-github

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.5 LTS (Noble Numbat)
AMD EPYC 9V74, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.403
  [Host]   : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  ShortRun : .NET 10.0.14 (10.0.1426.52326), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Count	200	152.2 μs	9.32 μs	0.51 μs	1.00	0.00	12.4512	2.1973	-	205.5 KB	1.00
Tfidf	200	168.3 μs	19.25 μs	1.06 μs	1.11	0.01	14.1602	3.4180	-	234.64 KB	1.14
CountBigrams	200	534.8 μs	55.51 μs	3.04 μs	3.51	0.02	35.1563	12.6953	-	586.99 KB	2.86
CountCharWordBoundary	200	1,679.6 μs	24.14 μs	1.32 μs	11.04	0.03	382.8125	382.8125	382.8125	1676.53 KB	8.16
Hashing	200	150.0 μs	9.68 μs	0.53 μs	0.99	0.00	8.3008	0.9766	-	136.13 KB	0.66
Count	1000	1,267.8 μs	51.66 μs	2.83 μs	1.00	0.00	109.3750	109.3750	109.3750	835.34 KB	1.00
Tfidf	1000	1,417.5 μs	280.36 μs	15.37 μs	1.12	0.01	142.5781	142.5781	142.5781	979.75 KB	1.17
CountBigrams	1000	3,085.2 μs	1,477.56 μs	80.99 μs	2.43	0.06	398.4375	398.4375	398.4375	2319.8 KB	2.78
CountCharWordBoundary	1000	10,970.5 μs	1,905.12 μs	104.43 μs	8.65	0.07	1046.8750	1015.6250	1015.6250	12013.04 KB	14.38
Hashing	1000	1,352.5 μs	571.10 μs	31.30 μs	1.07	0.02	70.3125	70.3125	70.3125	570.77 KB	0.68

Lodestar.Text.Benchmarks.VectorizerIncumbentBenchmarks-report-github

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

BenchmarkDotNet v0.14.0, Ubuntu 26.04.1 LTS (Resolute Raccoon)
AMD EPYC 7763, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.404
  [Host]   : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2
  ShortRun : .NET 10.0.15 (10.0.1526.53008), X64 RyuJIT AVX2

Job=ShortRun  IterationCount=3  LaunchCount=1  
WarmupCount=3

Method	Documents	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Lodestar	200	3.785 ms	1.020 ms	0.0559 ms	1.00	0.02	148.4375	148.4375	148.4375	1.78 MB	1.00
MlNet	200	54.625 ms	121.909 ms	6.6822 ms	14.43	1.54	7500.0000	7500.0000	7500.0000	28.27 MB	15.84
Lodestar	1000	17.714 ms	3.656 ms	0.2004 ms	1.00	0.01	1500.0000	1500.0000	1500.0000	8.11 MB	1.00
MlNet	1000	395.451 ms	302.671 ms	16.5904 ms	22.33	0.84	76000.0000	76000.0000	76000.0000	324.81 MB	40.06

compare-cdist

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'rapidfuzz': '3.14.6', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
cdist_ratio_n50	0.174	0.036	0.20x	0.174	0.036	0.20x
cdist_wratio_n50	2.415	2.272	0.94x	2.415	2.272	0.94x
cdist_ratio_varied_n50	0.136	0.031	0.23x	0.136	0.031	0.23x
cdist_ratio_varied_cutoff90_n50	0.056	0.031	0.55x	0.056	0.031	0.55x
cdist_ratio_n200	2.800	0.441	0.16x	2.804	0.441	0.16x
cdist_wratio_n200	36.958	34.476	0.93x	36.958	34.476	0.93x
cdist_ratio_varied_n200	2.367	0.388	0.16x	2.368	0.388	0.16x
cdist_ratio_varied_cutoff90_n200	1.016	0.388	0.38x	1.018	0.388	0.38x
cdist_ratio_n500	17.347	2.663	0.15x	17.366	2.663	0.15x
cdist_wratio_n500	232.313	209.204	0.90x	232.450	209.192	0.90x
cdist_ratio_varied_n500	14.135	2.296	0.16x	14.278	2.296	0.16x
cdist_ratio_varied_cutoff90_n500	5.751	2.300	0.40x	5.862	2.300	0.39x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-cluster-weighted

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'scikit-learn': '1.9.1', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
kmeans_weighted_10000	4.277	3.018	0.71x	4.282	6.033	1.41x
kmeans_restarts_10000	17.373	7.946	0.46x	17.397	15.858	0.91x
kmeans_weighted_100000	41.581	22.277	0.54x	42.016	39.605	0.94x
kmeans_restarts_100000	166.265	58.270	0.35x	166.834	109.896	0.66x
dbscan_weighted_5000	6.539	25.343	3.88x	6.978	25.342	3.63x
dbscan_weighted_20000	77.683	163.033	2.10x	79.190	163.024	2.06x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-cox-extended

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'lifelines': '0.30.3', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
stratified_weighted_1000	0.638	83.142	130.28x	0.638	83.137	130.28x
ridge_robust_1000	0.710	78.467	110.45x	0.710	78.465	110.45x
lasso_1000	5.197	716.112	137.80x	5.197	716.040	137.79x
ph_test_1000	0.097	14.619	151.24x	0.097	14.618	151.24x
predict_survival_1000	0.025	1.221	48.37x	0.025	1.221	48.37x
time_varying_1000	57.179	227.449	3.98x	57.175	227.423	3.98x
stratified_weighted_10000	7.346	466.554	63.51x	7.379	466.524	63.23x
ridge_robust_10000	8.160	805.342	98.69x	8.326	805.244	96.71x
lasso_10000	48.473	4727.613	97.53x	48.497	4727.453	97.48x
ph_test_10000	1.667	154.944	92.97x	1.720	154.939	90.06x
predict_survival_10000	0.027	1.225	45.55x	0.027	1.225	45.55x
time_varying_10000	2182.482	3753.357	1.72x	2182.393	3753.213	1.72x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-cross-conformal

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'mapie': '1.5.0', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
cv_plus_n1000	3.344	30.908	9.24x	3.344	122.966	36.77x
cv_minmax_n1000	2.563	4.421	1.72x	2.563	17.649	6.89x
cv_plus_gamma_n1000	3.941	31.489	7.99x	3.941	125.821	31.92x
bootstrap_plus_n1000	12.683	31.454	2.48x	12.685	125.554	9.90x
cv_plus_n10000	32.351	93.432	2.89x	32.351	372.948	11.53x
cv_minmax_n10000	27.137	20.361	0.75x	27.135	81.273	3.00x
cv_plus_gamma_n10000	47.534	101.201	2.13x	47.532	403.770	8.49x
bootstrap_plus_n10000	115.256	95.549	0.83x	115.249	380.968	3.31x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-glm

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'scipy': '1.18.1', 'statsmodels': '0.15.0', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
glm_negative_binomial_n1000	0.299	2.261	7.55x	0.299	2.261	7.55x
glm_gamma_n1000	0.359	3.021	8.42x	0.359	3.020	8.42x
glm_poisson_exposure_n1000	0.300	2.623	8.73x	0.300	2.623	8.73x
mnlogit_n1000	0.642	10.533	16.42x	0.641	10.532	16.42x
glm_negative_binomial_n10000	2.694	8.048	2.99x	2.700	8.047	2.98x
glm_gamma_n10000	3.045	9.494	3.12x	3.051	9.493	3.11x
glm_poisson_exposure_n10000	3.225	9.182	2.85x	3.231	9.182	2.84x
mnlogit_n10000	6.443	56.132	8.71x	6.448	56.127	8.70x
glm_negative_binomial_n100000	28.039	74.124	2.64x	28.493	295.747	10.38x
glm_gamma_n100000	32.202	87.638	2.72x	32.527	349.957	10.76x
glm_poisson_exposure_n100000	41.481	81.301	1.96x	42.087	324.676	7.71x
mnlogit_n100000	64.968	525.853	8.09x	65.004	1107.892	17.04x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-indel

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: rapidfuzz 3.14.6 (py 3.12.16)
C#:     Lodestar.Text on .NET 10.0.15 (mode Utf16Unit)

alphabet	length	Python ns/pair	C# ns/pair	speedup (py/C#)
latin	8	119.4	27.6	4.33x C# faster
latin	32	181.5	86.4	2.10x C# faster
latin	128	447.4	400.2	1.12x C# faster
latin	512	4465.9	4989.9	1.12x Py faster
cjk	8	145.3	25.8	5.63x C# faster
cjk	32	340.0	212.6	1.60x C# faster
cjk	128	1838.5	1395.1	1.32x C# faster
cjk	512	14917.0	9888.6	1.51x C# faster

Note: Indel is len(a)+len(b)-2*LCS on both sides, so this compares the subsequence kernels. Lodestar's is Hyyro's bit-parallel LLCS above a pattern of 8 and a rolling-row dynamic program below it (#273).

compare-iv

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'linearmodels': '7.0', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
2sls_robust_n1000	1.332	39.138	29.37x	1.332	39.133	29.37x
liml_robust_n1000	1.458	39.263	26.94x	1.457	39.260	26.94x
gmm_robust_n1000	1.342	38.801	28.91x	1.342	38.798	28.91x
2sls_kernel_n1000	7.295	41.115	5.64x	7.295	41.111	5.64x
2sls_clustered_n1000	1.294	40.349	31.18x	1.294	40.344	31.17x
2sls_robust_n10000	13.783	183.464	13.31x	14.119	366.700	25.97x
liml_robust_n10000	15.073	184.899	12.27x	15.474	369.641	23.89x
gmm_robust_n10000	13.778	178.419	12.95x	14.100	356.654	25.29x
2sls_kernel_n10000	77.891	192.617	2.47x	78.213	385.353	4.93x
2sls_clustered_n10000	13.191	199.319	15.11x	13.493	398.562	29.54x
2sls_robust_n100000	151.411	1947.008	12.86x	153.697	6361.652	41.39x
liml_robust_n100000	165.153	1974.396	11.95x	169.349	6441.831	38.04x
gmm_robust_n100000	153.234	1899.290	12.39x	155.692	6127.808	39.36x
2sls_kernel_n100000	2898.043	2716.147	0.94x	2899.524	8631.507	2.98x
2sls_clustered_n100000	143.446	2101.436	14.65x	146.786	6999.952	47.69x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-levenshtein

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: rapidfuzz 3.14.6 (py 3.12.16)
C#:     Lodestar.Text on .NET 10.0.15 (mode Utf16Unit)

alphabet	length	Python ns/pair	C# ns/pair	speedup (py/C#)
latin	8	149.9	18.1	8.30x C# faster
latin	32	282.0	159.2	1.77x C# faster
latin	128	2338.4	862.0	2.71x C# faster
latin	512	13987.0	11512.4	1.21x C# faster
cjk	8	159.8	17.9	8.93x C# faster
cjk	32	385.1	293.3	1.31x C# faster
cjk	128	3442.5	2095.8	1.64x C# faster
cjk	512	23736.0	19230.9	1.23x C# faster

Note: Python times the realistic per-call loop; rapidfuzz's C core uses the bit-parallel Myers algorithm, so it scales better on long strings.

compare-metrics

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'scikit-learn': '1.9.1', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
confusion_matrix_n1000_k2	0.005	0.961	212.97x	0.005	0.961	212.95x
accuracy_n1000_k2	0.000	0.512	4154.96x	0.000	0.512	4155.29x
precision_recall_f1_macro_n1000_k2	0.005	1.753	365.88x	0.005	1.753	365.89x
classification_report_n1000_k2	0.008	6.757	889.75x	0.008	6.757	889.76x
roc_auc_binary_n1000_k2	0.017	1.868	107.63x	0.017	1.868	107.60x
balanced_accuracy_n1000_k2	0.005	1.019	209.82x	0.005	1.019	209.82x
matthews_n1000_k2	0.005	1.953	424.65x	0.005	1.953	424.65x
cohen_kappa_n1000_k2	0.005	1.062	217.28x	0.005	1.062	217.30x
mse_n1000_k2	0.000	0.301	607.63x	0.000	0.301	607.63x
mae_n1000_k2	0.000	0.300	613.70x	0.000	0.299	613.69x
median_ae_n1000_k2	0.006	0.312	54.36x	0.006	0.312	54.36x
r2_n1000_k2	0.001	0.361	302.69x	0.001	0.361	302.67x
confusion_matrix_n1000_k10	0.005	0.957	186.70x	0.005	0.957	186.71x
accuracy_n1000_k10	0.000	0.509	4168.89x	0.000	0.509	4169.03x
precision_recall_f1_macro_n1000_k10	0.006	1.747	294.48x	0.006	1.747	294.51x
classification_report_n1000_k10	0.012	6.818	571.44x	0.012	6.817	571.42x
roc_auc_ovr_macro_n1000_k10	0.522	9.666	18.51x	0.522	9.666	18.51x
balanced_accuracy_n1000_k10	0.006	1.024	175.00x	0.006	1.024	175.00x
matthews_n1000_k10	0.006	1.971	347.51x	0.006	1.971	347.53x
cohen_kappa_n1000_k10	0.006	1.067	180.41x	0.006	1.067	180.40x
mse_n1000_k10	0.001	0.297	591.38x	0.001	0.297	591.44x
mae_n1000_k10	0.000	0.294	606.53x	0.000	0.294	606.53x
median_ae_n1000_k10	0.006	0.308	53.64x	0.006	0.308	53.64x
r2_n1000_k10	0.001	0.360	303.04x	0.001	0.360	303.02x
confusion_matrix_n100000_k2	0.476	10.732	22.54x	0.476	10.732	22.54x
accuracy_n100000_k2	0.012	3.755	305.17x	0.012	3.755	305.14x
precision_recall_f1_macro_n100000_k2	0.476	12.344	25.93x	0.476	12.343	25.92x
classification_report_n100000_k2	0.480	27.062	56.42x	0.480	27.060	56.42x
roc_auc_binary_n100000_k2	4.474	26.830	6.00x	4.474	26.828	6.00x
balanced_accuracy_n100000_k2	0.476	10.844	22.78x	0.476	10.843	22.78x
matthews_n100000_k2	0.476	21.627	45.47x	0.476	21.625	45.47x
cohen_kappa_n100000_k2	0.475	10.862	22.85x	0.475	10.861	22.85x
mse_n100000_k2	0.045	0.442	9.82x	0.045	0.441	9.82x
mae_n100000_k2	0.045	0.455	10.11x	0.045	0.454	10.11x
median_ae_n100000_k2	0.706	1.798	2.55x	0.735	1.798	2.44x
r2_n100000_k2	0.110	0.653	5.95x	0.110	0.653	5.95x
confusion_matrix_n100000_k10	0.508	10.809	21.28x	0.508	10.807	21.28x
accuracy_n100000_k10	0.012	3.763	306.39x	0.012	3.763	306.42x
precision_recall_f1_macro_n100000_k10	0.509	12.955	25.44x	0.509	12.955	25.44x
classification_report_n100000_k10	0.516	29.637	57.41x	0.516	29.636	57.41x
roc_auc_ovr_macro_n100000_k10	36.467	217.990	5.98x	36.464	217.979	5.98x
balanced_accuracy_n100000_k10	0.508	10.833	21.31x	0.508	10.833	21.31x
matthews_n100000_k10	0.508	22.665	44.57x	0.508	22.663	44.57x
cohen_kappa_n100000_k10	0.508	10.929	21.49x	0.509	10.929	21.49x
mse_n100000_k10	0.045	0.446	9.93x	0.045	0.446	9.93x
mae_n100000_k10	0.045	0.452	10.06x	0.045	0.452	10.06x
median_ae_n100000_k10	0.685	1.804	2.63x	0.731	1.804	2.47x
r2_n100000_k10	0.110	0.656	5.97x	0.110	0.656	5.97x
confusion_matrix_n1000000_k2	4.757	101.863	21.41x	4.757	101.821	21.40x
accuracy_n1000000_k2	0.135	33.508	248.29x	0.135	33.496	248.22x
precision_recall_f1_macro_n1000000_k2	4.765	109.329	22.94x	4.765	109.321	22.94x
classification_report_n1000000_k2	4.770	213.874	44.84x	4.770	213.843	44.83x
roc_auc_binary_n1000000_k2	58.972	296.864	5.03x	58.970	296.847	5.03x
balanced_accuracy_n1000000_k2	4.759	101.249	21.28x	4.759	101.237	21.27x
matthews_n1000000_k2	4.758	202.576	42.57x	4.758	202.550	42.57x
cohen_kappa_n1000000_k2	4.763	101.296	21.27x	4.763	101.291	21.27x
mse_n1000000_k2	0.459	1.882	4.10x	0.459	1.880	4.10x
mae_n1000000_k2	0.458	2.460	5.37x	0.458	2.460	5.37x
median_ae_n1000000_k2	7.495	14.577	1.94x	7.716	14.577	1.89x
r2_n1000000_k2	1.116	3.137	2.81x	1.116	3.131	2.81x
confusion_matrix_n1000000_k10	5.076	101.702	20.03x	5.076	101.529	20.00x
accuracy_n1000000_k10	0.132	33.928	257.16x	0.132	33.923	257.14x
precision_recall_f1_macro_n1000000_k10	5.077	116.280	22.90x	5.078	116.273	22.90x
classification_report_n1000000_k10	5.084	239.941	47.19x	5.084	239.927	47.19x
balanced_accuracy_n1000000_k10	5.075	102.269	20.15x	5.075	102.265	20.15x
matthews_n1000000_k10	5.073	211.202	41.63x	5.073	211.177	41.63x
cohen_kappa_n1000000_k10	5.070	102.084	20.14x	5.070	102.080	20.13x
mse_n1000000_k10	0.464	2.135	4.61x	0.464	2.131	4.60x
mae_n1000000_k10	0.465	2.526	5.43x	0.465	2.526	5.43x
median_ae_n1000000_k10	6.418	14.682	2.29x	6.473	14.681	2.27x
r2_n1000000_k10	1.142	3.389	2.97x	1.142	3.387	2.97x

ratio > 1 means Lodestar is faster. cpu is the merge gate for this branch (docs/guides/performance.md): every operation, every size, must be >= 1x.

compare-nmf-transform

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'scikit-learn': '1.9.1', 'scipy': '1.18.1'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
transform_frobenius_n100	2.463	0.867	0.35x	2.463	0.867	0.35x
transform_frobenius_n500	7.818	1.683	0.22x	7.818	1.683	0.22x
transform_frobenius_n2000	28.345	5.814	0.21x	28.387	23.048	0.81x
transform_kl_n100	1.131	17.653	15.61x	1.131	17.653	15.61x
transform_kl_n500	5.082	34.623	6.81x	5.082	34.622	6.81x
transform_kl_n2000	20.274	123.314	6.08x	20.361	448.869	22.05x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-ols

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'scipy': '1.18.1', 'statsmodels': '0.15.0', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
ols_summary_n1000	0.071	1.929	27.06x	0.071	1.929	27.06x
ols_hac_n1000	0.134	2.370	17.62x	0.134	2.369	17.62x
ols_cluster_n1000	0.051	2.422	47.94x	0.051	2.422	47.94x
ols_summary_n10000	0.379	8.838	23.34x	0.379	8.837	23.33x
ols_hac_n10000	1.381	9.769	7.08x	1.386	9.769	7.05x
ols_cluster_n10000	0.499	10.101	20.22x	0.502	10.100	20.12x
ols_summary_n100000	3.677	82.727	22.50x	3.692	330.394	89.48x
ols_hac_n100000	14.359	92.827	6.46x	14.691	370.828	25.24x
ols_cluster_n100000	5.768	94.871	16.45x	6.005	378.571	63.05x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-onehot

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'scikit-learn': '1.9.1', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
sparse_10000	1.413	3.352	2.37x	1.417	3.352	2.37x
sparse_infrequent_10000	1.784	4.107	2.30x	1.788	4.107	2.30x
sparse_100000	14.248	24.381	1.71x	14.552	24.379	1.68x
sparse_infrequent_100000	16.078	29.179	1.81x	16.404	29.178	1.78x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-panel

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'linearmodels': '7.0', 'numpy': '2.5.3', 'pandas': '2.3.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
fe_entity_n1000	0.202	13.780	68.15x	0.202	13.779	68.15x
fe_twoway_clustered_n1000	0.333	21.522	64.59x	0.333	21.522	64.60x
fe_kernel_n1000	0.215	14.156	65.81x	0.215	14.156	65.81x
between_robust_n1000	0.121	13.313	109.90x	0.121	13.312	109.90x
first_difference_robust_n1000	0.229	15.760	68.69x	0.229	15.760	68.69x
random_effects_n1000	0.213	18.187	85.23x	0.213	18.186	85.23x
fe_entity_n10000	2.146	21.354	9.95x	2.218	21.353	9.63x
fe_twoway_clustered_n10000	3.867	41.167	10.65x	4.085	82.352	20.16x
fe_kernel_n10000	2.335	21.801	9.33x	2.362	21.799	9.23x
between_robust_n10000	1.454	20.591	14.16x	1.457	20.590	14.14x
first_difference_robust_n10000	2.394	24.601	10.27x	2.400	24.600	10.25x
random_effects_n10000	2.265	28.450	12.56x	2.294	28.448	12.40x
fe_entity_n100000	28.247	134.121	4.75x	30.043	535.679	17.83x
fe_twoway_clustered_n100000	41.070	262.730	6.40x	43.152	1010.587	23.42x
fe_kernel_n100000	27.800	124.186	4.47x	28.451	495.739	17.42x
between_robust_n100000	15.826	126.249	7.98x	16.473	465.417	28.25x
first_difference_robust_n100000	27.426	142.535	5.20x	28.717	569.488	19.83x
random_effects_n100000	27.162	167.598	6.17x	28.556	669.352	23.44x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-persistence

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'tokenizers': '0.23.3', 'sentencepiece': '0.2.3', 'scikit-learn': '1.9.1', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu	C# bytes	Py bytes
vocab_txt	5.652	9.731	1.72x	5.844	9.731	1.67x	228,891	228,891
tokenizer_json_wordpiece	12.520	17.246	1.38x	12.902	17.245	1.34x	706,526	706,526
tokenizer_json_unigram	13.712	41.528	3.03x	13.859	41.524	3.00x	1,990,038	1,990,038
spiece_model	6.340	15.088	2.38x	6.611	15.087	2.28x	347,061	347,061
tfidf_save	1.611	2.513	1.56x	1.653	2.513	1.52x	581,787	591,922
tfidf_load	4.963	4.134	0.83x	5.230	4.134	0.79x	581,787	591,922
embedding_index_save	5.003	1.775	0.35x	5.193	1.775	0.34x	20,589,007	15,360,128
embedding_index_save_file	49.455	37.484	0.76x	11.353	5.176	0.46x	20,589,007	15,360,128
embedding_index_load	7.835	1.640	0.21x	8.305	1.635	0.20x	20,589,007	15,360,128
embedding_index_load_file	7.786	1.058	0.14x	8.332	1.058	0.13x	20,589,007	15,360,128
embedding_index_load_memory	6.371	1.645	0.26x	6.747	1.636	0.24x	20,589,007	15,360,128
embedding_index_ingest_npy	2.682	1.596	0.60x	3.020	1.596	0.53x	15,360,128	15,360,128
embedding_index_view_floor	0.000	0.001	88.20x	0.000	0.001	88.20x	20,589,007	15,360,128
embedding_index_save_gzip	413.694	584.562	1.41x	416.589	584.540	1.40x	15,250,490	14,022,374
embedding_index_load_gzip	73.984	66.942	0.90x	74.024	66.942	0.90x	15,250,490	14,022,374

ratio > 1 means Lodestar is faster. cpu is the honest one: elapsed time hides work .NET does on background GC threads; CPython is single-threaded. bytes is what the row wrote or read; a results file from before #378 carries none, and the two columns then disappear rather than read zero.

compare-splitters

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'scikit-learn': '1.9.1', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
kfold_n10000	0.092	0.113	1.22x	0.092	0.113	1.22x
stratified_n10000	0.188	0.859	4.56x	0.188	0.859	4.56x
traintest_n10000	0.005	0.234	47.42x	0.005	0.234	47.41x
kfold_seeded_n10000	0.307	0.655	2.13x	0.307	0.655	2.13x
stratified_seeded_n10000	0.438	1.436	3.28x	0.438	1.436	3.28x
traintest_seeded_n10000	0.165	0.497	3.02x	0.165	0.497	3.02x
stratified_traintest_n10000	0.667	2.672	4.01x	0.667	2.672	4.01x
group_kfold_n10000	0.122	0.580	4.74x	0.122	0.580	4.74x
stratified_group_kfold_n10000	0.241	30.769	127.89x	0.241	30.766	127.89x
timeseries_n10000	0.000	0.012	86.54x	0.000	0.012	86.54x
repeated_kfold_n10000	0.992	1.758	1.77x	0.992	1.758	1.77x
kfold_n100000	1.196	2.433	2.03x	1.359	2.433	1.79x
stratified_n100000	2.174	7.139	3.28x	2.337	7.119	3.05x
traintest_n100000	0.222	0.399	1.80x	0.224	0.399	1.78x
kfold_seeded_n100000	3.591	7.267	2.02x	3.781	7.251	1.92x
stratified_seeded_n100000	4.811	12.586	2.62x	5.066	12.569	2.48x
traintest_seeded_n100000	2.030	2.003	0.99x	2.077	2.003	0.96x
stratified_traintest_n100000	8.090	23.733	2.93x	8.204	23.715	2.89x
group_kfold_n100000	1.645	7.111	4.32x	1.824	7.092	3.89x
stratified_group_kfold_n100000	2.772	308.106	111.15x	2.962	308.082	104.02x
timeseries_n100000	0.000	0.043	324.05x	0.000	0.043	324.11x
repeated_kfold_n100000	10.436	21.091	2.02x	10.630	21.090	1.98x
kfold_n1000000	11.468	20.720	1.81x	12.372	20.718	1.67x
stratified_n1000000	23.514	67.689	2.88x	24.962	67.515	2.70x
traintest_n1000000	0.803	2.116	2.64x	0.948	2.109	2.22x
kfold_seeded_n1000000	39.483	62.558	1.58x	41.193	62.482	1.52x
stratified_seeded_n1000000	52.596	109.271	2.08x	54.643	109.266	2.00x
traintest_seeded_n1000000	20.730	19.196	0.93x	21.136	19.195	0.91x
stratified_traintest_n1000000	93.846	235.609	2.51x	94.361	235.600	2.50x
group_kfold_n1000000	18.579	57.014	3.07x	19.864	56.973	2.87x
stratified_group_kfold_n1000000	31.294	3081.712	98.48x	33.263	3079.309	92.57x
timeseries_n1000000	0.000	0.328	2925.63x	0.000	0.328	2925.82x
repeated_kfold_n1000000	112.376	182.911	1.63x	115.079	182.906	1.59x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-stats

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'scipy': '1.18.1', 'statsmodels': '0.15.0', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
welch_t_n1000	0.008	0.803	100.46x	0.008	0.803	100.43x
mann_whitney_n1000	0.030	0.723	23.79x	0.030	0.723	23.79x
chi_square_n1000	0.000	0.332	1657.56x	0.000	0.332	1657.50x
fligner_n1000	0.073	0.796	10.83x	0.073	0.796	10.84x
anderson_ksamp_n1000	0.164	0.393	2.40x	0.164	0.393	2.40x
spearman_matrix_n1000	0.128	0.910	7.10x	0.128	0.910	7.10x
pointbiserial_n1000	0.010	0.566	54.09x	0.010	0.566	54.08x
welch_t_n10000	0.053	0.853	15.98x	0.053	0.853	15.98x
mann_whitney_n10000	0.682	2.803	4.11x	0.682	2.803	4.11x
chi_square_n10000	0.001	0.335	239.97x	0.001	0.335	239.93x
fligner_n10000	1.443	3.228	2.24x	1.533	3.228	2.11x
anderson_ksamp_n10000	2.054	4.025	1.96x	2.188	4.025	1.84x
spearman_matrix_n10000	2.149	4.774	2.22x	2.215	4.774	2.16x
pointbiserial_n10000	0.100	0.622	6.21x	0.100	0.622	6.21x
welch_t_n100000	0.518	1.245	2.40x	0.518	1.245	2.40x
mann_whitney_n100000	7.746	28.891	3.73x	7.746	28.889	3.73x
chi_square_n100000	0.014	0.360	25.70x	0.014	0.360	25.70x
fligner_n100000	14.892	33.398	2.24x	15.345	33.396	2.18x
anderson_ksamp_n100000	15.780	43.762	2.77x	16.185	43.760	2.70x
spearman_matrix_n100000	19.245	54.632	2.84x	19.831	54.629	2.75x
pointbiserial_n100000	1.300	1.380	1.06x	1.318	5.506	4.18x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-survival-aalen

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'lifelines': '0.30.3', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
fit_1000	6.883	49.532	7.20x	6.883	197.417	28.68x
fit_penalised_weighted_1000	6.572	53.731	8.18x	6.575	214.536	32.63x
predict_survival_1000	0.717	1.455	2.03x	0.719	1.455	2.02x
fit_10000	195.772	498.089	2.54x	196.025	846.591	4.32x
fit_penalised_weighted_10000	183.611	511.471	2.79x	183.855	857.973	4.67x
predict_survival_10000	2.888	2.008	0.70x	3.074	8.002	2.60x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-survival-family

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'lifelines': '0.30.3', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
logrank_wilcoxon_weighted_1000	0.019	9.293	486.18x	0.019	9.292	486.15x
logrank_fleming_harrington_1000	0.027	14.106	530.59x	0.027	14.105	530.57x
multigroup_1000	0.090	19.155	212.69x	0.090	19.153	212.68x
pairwise_1000	0.223	97.844	439.29x	0.223	97.839	439.27x
rmst_1000	0.072	803.353	11214.44x	0.072	803.313	11214.47x
fixed_point_1000	0.029	10.117	349.10x	0.029	10.116	349.09x
concordance_1000	0.040	6.031	149.17x	0.040	6.031	149.16x
logrank_wilcoxon_weighted_10000	0.203	10.790	53.27x	0.203	10.789	53.28x
logrank_fleming_harrington_10000	0.275	17.548	63.80x	0.275	17.547	63.79x
multigroup_10000	2.053	29.556	14.40x	2.059	29.555	14.35x
pairwise_10000	3.833	112.503	29.35x	3.833	112.498	29.35x
rmst_10000	1.536	868.429	565.22x	1.567	868.408	554.27x
fixed_point_10000	0.296	12.202	41.29x	0.296	12.202	41.28x
concordance_10000	2.049	72.710	35.49x	2.049	72.708	35.49x
logrank_wilcoxon_weighted_100000	3.715	18.338	4.94x	3.803	18.338	4.82x
logrank_fleming_harrington_100000	3.680	28.899	7.85x	3.731	28.897	7.74x
multigroup_100000	9.555	41.391	4.33x	9.734	41.390	4.25x
pairwise_100000	38.409	185.694	4.83x	39.473	185.686	4.70x
rmst_100000	8.060	322.456	40.01x	8.149	322.447	39.57x
fixed_point_100000	4.151	22.953	5.53x	4.276	22.952	5.37x
concordance_100000	25.555	603.202	23.60x	25.732	603.172	23.44x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-survival-parametric

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'lifelines': '0.30.3', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
weibull_1000	2.656	45.665	17.19x	2.656	45.663	17.19x
loglogistic_left_1000	4.748	65.535	13.80x	4.747	65.531	13.80x
lognormal_interval_1000	4.271	118.058	27.64x	4.271	118.053	27.64x
generalized_gamma_1000	32.904	641.485	19.50x	32.959	641.442	19.46x
weibull_aft_1000	5.305	85.961	16.20x	5.330	85.959	16.13x
lognormal_aft_ridge_robust_1000	8.395	3128.746	372.68x	8.471	3128.504	369.31x
loglogistic_aft_interval_1000	8.209	135.271	16.48x	8.211	135.257	16.47x
breslow_fleming_harrington_1000	0.114	4.673	41.15x	0.114	4.673	41.15x
kaplan_meier_left_1000	0.069	5.495	79.32x	0.069	5.495	79.31x
weibull_10000	26.185	51.739	1.98x	26.181	51.738	1.98x
loglogistic_left_10000	52.726	108.342	2.05x	52.723	108.333	2.05x
lognormal_interval_10000	43.531	221.341	5.08x	43.527	221.324	5.08x
generalized_gamma_10000	334.091	1596.846	4.78x	334.074	1596.794	4.78x
weibull_aft_10000	53.776	96.514	1.79x	53.772	96.509	1.79x
lognormal_aft_ridge_robust_10000	108.621	29990.667	276.10x	108.619	29989.419	276.10x
loglogistic_aft_interval_10000	77.387	179.638	2.32x	77.390	179.624	2.32x
breslow_fleming_harrington_10000	2.121	6.146	2.90x	2.121	6.146	2.90x
kaplan_meier_left_10000	0.836	7.225	8.65x	0.836	7.225	8.65x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-transformers

As of 2026-11-11, measured at commit 59a57f7257ac13a8a57dcee55e218a7f45b3b458.

Python: {'scikit-learn': '1.9.1', 'numpy': '2.5.4'} (py 3.12.16)
C#:     Lodestar on .NET 10.0.15

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
normalize_n10000	0.282	0.457	1.62x	0.283	0.457	1.61x
polynomial_n10000	0.778	1.015	1.30x	0.785	1.015	1.29x
kbins_fit_n10000	1.469	2.592	1.76x	1.469	2.592	1.76x
kbins_n10000	0.740	4.723	6.39x	0.853	4.723	5.54x
quantile_fit_n10000	1.568	3.807	2.43x	1.569	3.807	2.43x
quantile_n10000	1.843	1.708	0.93x	1.847	1.708	0.92x
power_fit_n10000	5.919	128.264	21.67x	5.923	128.256	21.65x
power_n10000	1.568	1.434	0.91x	1.571	1.434	0.91x
knn_impute_n10000	7.326	25.315	3.46x	7.326	99.906	13.64x
label_n10000	1.162	0.775	0.67x	1.162	0.775	0.67x
normalize_n100000	1.578	1.925	1.22x	2.018	1.925	0.95x
polynomial_n100000	3.910	7.544	1.93x	4.770	7.542	1.58x
kbins_fit_n100000	15.083	8.354	0.55x	15.961	8.353	0.52x
kbins_n100000	8.030	13.231	1.65x	9.053	13.230	1.46x
quantile_fit_n100000	15.430	18.782	1.22x	16.223	18.780	1.16x
quantile_n100000	16.941	15.557	0.92x	17.429	15.556	0.89x
power_fit_n100000	63.743	873.707	13.71x	66.698	873.675	13.10x
power_n100000	14.491	9.648	0.67x	14.993	9.647	0.64x
knn_impute_n100000	7.316	19.632	2.68x	7.316	78.356	10.71x
label_n100000	15.513	7.425	0.48x	15.810	7.425	0.47x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.

compare-var

As of 2026-11-08, measured at commit fd0cf96a1a303125ff7b6b7941a2ca17e5613a1c.

Python: {'scipy': '1.18.1', 'statsmodels': '0.15.0', 'numpy': '2.5.3'} (py 3.12.15)
C#:     Lodestar on .NET 10.0.14

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
var_n1000	0.032	1.933	61.08x	0.032	1.933	61.08x
var_n10000	0.381	12.453	32.69x	0.383	12.452	32.52x
var_n100000	3.156	118.393	37.51x	3.287	118.385	36.02x

ratio > 1 means Lodestar is faster. cpu is the honest one on both sides: elapsed time hides .NET's background GC threads, and numpy's BLAS threads too -- statsmodels at 100 000 rows spends ~10x more cpu than wall.