benchmark_latest

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.Text.Benchmarks.BatchEmbeddingBenchmarks-report-github

As of 2026-08-20, measured at commit 6cadb6a76b4e172c36be7cdb3f541d742e24373a.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	CorpusSize	Mean	Error	StdDev	Ratio	Gen0	Gen1	Allocated	Alloc Ratio
UnitLoop	1	6.147 μs	0.4325 μs	0.0237 μs	1.00	0.1678	-	2.76 KB	1.00
EmbedBatch	1	6.426 μs	1.1243 μs	0.0616 μs	1.05	0.1907	-	3.16 KB	1.14
EmbedBatchBucketed	1	6.512 μs	0.3421 μs	0.0187 μs	1.06	0.1907	-	3.16 KB	1.14
UnitLoop	8	111.178 μs	1.2248 μs	0.0671 μs	1.00	8.1787	0.2441	134.29 KB	1.00
EmbedBatch	8	74.442 μs	14.5132 μs	0.7955 μs	0.67	7.6904	0.3662	127.31 KB	0.95
EmbedBatchBucketed	8	73.575 μs	5.0715 μs	0.2780 μs	0.66	7.6904	0.3662	127.31 KB	0.95
UnitLoop	32	412.764 μs	26.4858 μs	1.4518 μs	1.00	28.3203	0.9766	468.71 KB	1.00
EmbedBatch	32	257.827 μs	59.9502 μs	3.2861 μs	0.62	26.8555	1.4648	441.32 KB	0.94
EmbedBatchBucketed	32	250.795 μs	24.6911 μs	1.3534 μs	0.61	25.8789	1.4648	427.8 KB	0.91
UnitLoop	128	1,662.869 μs	140.3652 μs	7.6939 μs	1.00	113.2813	5.8594	1874.78 KB	1.00
EmbedBatch	128	1,023.228 μs	71.7800 μs	3.9345 μs	0.62	107.4219	15.6250	1764.42 KB	0.94
EmbedBatchBucketed	128	946.209 μs	47.8914 μs	2.6251 μs	0.57	103.5156	15.6250	1696.9 KB	0.91

Lodestar.Text.Benchmarks.BpeBenchmarks-report-github

As of 2026-08-20, measured at commit 6cadb6a76b4e172c36be7cdb3f541d742e24373a.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Mean	Error	StdDev	Ratio	Gen0	Allocated	Alloc Ratio
Unigram	607.0 ms	49.79 ms	2.73 ms	1.00	32000.0000	519.51 MB	1.00
Bpe	559.4 ms	46.95 ms	2.57 ms	0.92	7000.0000	112.18 MB	0.22

Lodestar.Text.Benchmarks.BpeScalingBenchmarks-report-github

As of 2026-08-20, measured at commit 6cadb6a76b4e172c36be7cdb3f541d742e24373a.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Gen0	Allocated
BpeOnOnePathologicalToken	512	103.7 μs	2.49 μs	0.14 μs	1.2207	20.38 KB
BpeOnOnePathologicalToken	1024	221.4 μs	32.67 μs	1.79 μs	2.4414	39.93 KB
BpeOnOnePathologicalToken	2048	483.9 μs	27.12 μs	1.49 μs	3.9063	78.98 KB
BpeOnOnePathologicalToken	4096	987.9 μs	41.70 μs	2.29 μs	7.8125	157.03 KB

Lodestar.Text.Benchmarks.BucketRouteDiagnostics-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Mean	Error	StdDev	Allocated
DpGroup	13.66 μs	0.293 μs	0.016 μs	-
MyersGroup	114.45 μs	1.663 μs	0.091 μs	-

Lodestar.Text.Benchmarks.FuzzBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
Ratio	108.8 ns	1.56 ns	0.09 ns	1.00	0.00	-	-	NA
PartialRatio	18,009.4 ns	903.43 ns	49.52 ns	165.48	0.41	-	-	NA
TokenSortRatio	944.5 ns	39.15 ns	2.15 ns	8.68	0.02	0.0782	1312 B	NA
TokenSetRatio	3,184.8 ns	132.87 ns	7.28 ns	29.26	0.06	0.3433	5760 B	NA
WRatio	4,565.5 ns	850.73 ns	46.63 ns	41.95	0.37	0.4272	7200 B	NA

Lodestar.Text.Benchmarks.IndelBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_Utf16	8	29.47 ns	0.283 ns	0.015 ns	1.00	0.00	-	NA
Distance_CodePoint	8	131.49 ns	7.121 ns	0.390 ns	4.46	0.01	-	NA
NormalizedSimilarity_Utf16	8	30.27 ns	0.552 ns	0.030 ns	1.03	0.00	-	NA
SubsequenceLength_Utf16	8	28.81 ns	0.194 ns	0.011 ns	0.98	0.00	-	NA
Distance_Utf16	12	32.06 ns	0.542 ns	0.030 ns	1.00	0.00	-	NA
Distance_CodePoint	12	139.71 ns	4.418 ns	0.242 ns	4.36	0.01	-	NA
NormalizedSimilarity_Utf16	12	32.95 ns	0.115 ns	0.006 ns	1.03	0.00	-	NA
SubsequenceLength_Utf16	12	30.50 ns	0.998 ns	0.055 ns	0.95	0.00	-	NA
Distance_Utf16	16	34.77 ns	1.635 ns	0.090 ns	1.00	0.00	-	NA
Distance_CodePoint	16	147.94 ns	22.131 ns	1.213 ns	4.25	0.03	-	NA
NormalizedSimilarity_Utf16	16	36.70 ns	1.701 ns	0.093 ns	1.06	0.00	-	NA
SubsequenceLength_Utf16	16	35.67 ns	1.469 ns	0.081 ns	1.03	0.00	-	NA
Distance_Utf16	20	89.91 ns	1.795 ns	0.098 ns	1.00	0.00	-	NA
Distance_CodePoint	20	186.54 ns	2.142 ns	0.117 ns	2.07	0.00	-	NA
NormalizedSimilarity_Utf16	20	91.64 ns	1.835 ns	0.101 ns	1.02	0.00	-	NA
SubsequenceLength_Utf16	20	92.23 ns	2.542 ns	0.139 ns	1.03	0.00	-	NA
Distance_Utf16	24	61.77 ns	3.992 ns	0.219 ns	1.00	0.00	-	NA
Distance_CodePoint	24	859.29 ns	352.191 ns	19.305 ns	13.91	0.27	-	NA
NormalizedSimilarity_Utf16	24	57.71 ns	8.947 ns	0.490 ns	0.93	0.01	-	NA
SubsequenceLength_Utf16	24	63.84 ns	10.009 ns	0.549 ns	1.03	0.01	-	NA
Distance_Utf16	32	65.22 ns	3.411 ns	0.187 ns	1.00	0.00	-	NA
Distance_CodePoint	32	1,063.72 ns	207.056 ns	11.349 ns	16.31	0.16	-	NA
NormalizedSimilarity_Utf16	32	69.71 ns	0.601 ns	0.033 ns	1.07	0.00	-	NA
SubsequenceLength_Utf16	32	65.01 ns	3.627 ns	0.199 ns	1.00	0.00	-	NA
Distance_Utf16	128	1,449.13 ns	155.845 ns	8.542 ns	1.00	0.01	-	NA
Distance_CodePoint	128	21,824.95 ns	22,364.650 ns	1,225.882 ns	15.06	0.74	-	NA
NormalizedSimilarity_Utf16	128	1,465.74 ns	83.533 ns	4.579 ns	1.01	0.01	-	NA
SubsequenceLength_Utf16	128	1,408.28 ns	5.026 ns	0.276 ns	0.97	0.00	-	NA
Distance_Utf16	512	12,544.34 ns	219.591 ns	12.037 ns	1.00	0.00	-	NA
Distance_CodePoint	512	310,332.47 ns	77,998.119 ns	4,275.340 ns	24.74	0.30	-	NA
NormalizedSimilarity_Utf16	512	12,625.76 ns	81.047 ns	4.442 ns	1.01	0.00	-	NA
SubsequenceLength_Utf16	512	12,663.70 ns	92.240 ns	5.056 ns	1.01	0.00	-	NA

Lodestar.Text.Benchmarks.LcsGateBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Band	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Dp	8	130.92 ns	29.556 ns	1.620 ns	1.00	0.02	-	NA
Kernel	8	65.76 ns	1.471 ns	0.081 ns	0.50	0.01	-	NA
Dp	12	222.28 ns	39.802 ns	2.182 ns	1.00	0.01	-	NA
Kernel	12	69.46 ns	0.501 ns	0.027 ns	0.31	0.00	-	NA
Dp	14	296.74 ns	322.353 ns	17.669 ns	1.00	0.07	-	NA
Kernel	14	73.20 ns	0.596 ns	0.033 ns	0.25	0.01	-	NA
Dp	16	358.57 ns	299.625 ns	16.423 ns	1.00	0.06	-	NA
Kernel	16	77.81 ns	0.883 ns	0.048 ns	0.22	0.01	-	NA
Dp	18	745.92 ns	301.683 ns	16.536 ns	1.00	0.03	-	NA
Kernel	18	83.55 ns	3.556 ns	0.195 ns	0.11	0.00	-	NA
Dp	20	828.44 ns	140.667 ns	7.710 ns	1.00	0.01	-	NA
Kernel	20	88.93 ns	0.835 ns	0.046 ns	0.11	0.00	-	NA
Dp	24	1,005.43 ns	323.306 ns	17.721 ns	1.00	0.02	-	NA
Kernel	24	96.48 ns	1.030 ns	0.056 ns	0.10	0.00	-	NA
Dp	32	1,603.80 ns	721.131 ns	39.528 ns	1.00	0.03	-	NA
Kernel	32	116.79 ns	27.889 ns	1.529 ns	0.07	0.00	-	NA
Dp	48	3,259.76 ns	713.374 ns	39.102 ns	1.00	0.01	-	NA
Kernel	48	145.96 ns	17.750 ns	0.973 ns	0.04	0.00	-	NA
Dp	64	5,258.48 ns	3,832.506 ns	210.073 ns	1.00	0.05	-	NA
Kernel	64	173.82 ns	2.862 ns	0.157 ns	0.03	0.00	-	NA
Dp	96	11,607.35 ns	2,164.185 ns	118.626 ns	1.00	0.01	-	NA
Kernel	96	1,146.96 ns	15.852 ns	0.869 ns	0.10	0.00	-	NA

Lodestar.Text.Benchmarks.LevenshteinBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_Utf16	8	34.56 ns	1.854 ns	0.102 ns	1.00	0.00	-	NA
Distance_CodePoint	8	128.73 ns	42.829 ns	2.348 ns	3.73	0.06	-	NA
NormalizedSimilarity_Utf16	8	28.12 ns	0.211 ns	0.012 ns	0.81	0.00	-	NA
Distance_Utf16	64	304.54 ns	109.428 ns	5.998 ns	1.00	0.02	-	NA
Distance_CodePoint	64	738.81 ns	196.952 ns	10.796 ns	2.43	0.05	-	NA
NormalizedSimilarity_Utf16	64	303.71 ns	4.114 ns	0.225 ns	1.00	0.02	-	NA
Distance_Utf16	512	15,337.55 ns	673.909 ns	36.939 ns	1.00	0.00	-	NA
Distance_CodePoint	512	18,758.97 ns	2,026.986 ns	111.106 ns	1.22	0.01	-	NA
NormalizedSimilarity_Utf16	512	16,412.86 ns	3,675.965 ns	201.492 ns	1.07	0.01	-	NA

Lodestar.Text.Benchmarks.LevenshteinCodePointBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Distinct	Mean	Error	StdDev	Ratio	RatioSD	Allocated	Alloc Ratio
Distance_CodePoint	16	32	354.1 ns	2.73 ns	0.15 ns	1.00	0.00	-	NA
Distance_Utf16	16	32	1,516.3 ns	40.90 ns	2.24 ns	4.28	0.01	-	NA
Distance_CodePoint	16	512	353.0 ns	0.59 ns	0.03 ns	1.00	0.00	-	NA
Distance_Utf16	16	512	1,513.7 ns	205.49 ns	11.26 ns	4.29	0.03	-	NA
Distance_CodePoint	24	32	446.4 ns	4.23 ns	0.23 ns	1.00	0.00	-	NA
Distance_Utf16	24	32	3,423.2 ns	42.93 ns	2.35 ns	7.67	0.01	-	NA
Distance_CodePoint	24	512	489.7 ns	389.40 ns	21.34 ns	1.00	0.05	-	NA
Distance_Utf16	24	512	3,557.7 ns	109.33 ns	5.99 ns	7.27	0.28	-	NA
Distance_CodePoint	32	32	535.4 ns	0.32 ns	0.02 ns	1.00	0.00	-	NA
Distance_Utf16	32	32	6,063.4 ns	601.71 ns	32.98 ns	11.33	0.05	-	NA
Distance_CodePoint	32	512	566.7 ns	14.74 ns	0.81 ns	1.00	0.00	-	NA
Distance_Utf16	32	512	6,330.7 ns	251.83 ns	13.80 ns	11.17	0.03	-	NA
Distance_CodePoint	40	32	623.9 ns	47.05 ns	2.58 ns	1.00	0.01	-	NA
Distance_Utf16	40	32	10,211.9 ns	223.72 ns	12.26 ns	16.37	0.06	-	NA
Distance_CodePoint	40	512	654.8 ns	68.17 ns	3.74 ns	1.00	0.01	-	NA
Distance_Utf16	40	512	9,799.3 ns	141.43 ns	7.75 ns	14.97	0.07	-	NA
Distance_CodePoint	128	32	2,384.3 ns	56.06 ns	3.07 ns	1.00	0.00	-	NA
Distance_Utf16	128	32	103,097.9 ns	15,093.75 ns	827.34 ns	43.24	0.30	-	NA
Distance_CodePoint	128	512	2,406.0 ns	8.50 ns	0.47 ns	1.00	0.00	-	NA
Distance_Utf16	128	512	109,957.4 ns	15,676.07 ns	859.26 ns	45.70	0.31	-	NA
Distance_CodePoint	512	32	20,475.5 ns	169.41 ns	9.29 ns	1.00	0.00	-	NA
Distance_Utf16	512	32	2,391,970.0 ns	174,601.66 ns	9,570.51 ns	116.82	0.41	-	NA
Distance_CodePoint	512	512	447,403.6 ns	69,150.90 ns	3,790.39 ns	1.00	0.01	-	NA
Distance_Utf16	512	512	1,803,494.5 ns	86,631.99 ns	4,748.59 ns	4.03	0.03	-	NA

Lodestar.Text.Benchmarks.MetricsBenchmarks-report-github

As of 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Samples	Classes	Mean	Error	StdDev	Gen0	Allocated
Matrix	1000	2	7,805.4 ns	355.41 ns	19.48 ns	0.0153	312 B
MatrixWeighted	1000	2	7,191.7 ns	142.88 ns	7.83 ns	0.0153	312 B
AccuracyScore	1000	2	921.0 ns	26.54 ns	1.45 ns	-	-
F1Macro	1000	2	7,638.9 ns	540.01 ns	29.60 ns	0.0153	472 B
Report	1000	2	10,871.3 ns	2,905.72 ns	159.27 ns	0.3815	6520 B
Matrix	1000	10	7,886.8 ns	945.29 ns	51.81 ns	0.0610	1248 B
MatrixWeighted	1000	10	7,493.1 ns	117.68 ns	6.45 ns	0.0687	1248 B
AccuracyScore	1000	10	918.7 ns	24.97 ns	1.37 ns	-	-
F1Macro	1000	10	8,066.8 ns	403.99 ns	22.14 ns	0.0916	1664 B
Report	1000	10	15,579.9 ns	3,045.75 ns	166.95 ns	0.9155	15496 B
Matrix	100000	2	873,748.0 ns	183,124.81 ns	10,037.69 ns	-	313 B
MatrixWeighted	100000	2	828,270.5 ns	28,685.16 ns	1,572.33 ns	-	313 B
AccuracyScore	100000	2	163,520.3 ns	4,952.32 ns	271.45 ns	-	-
F1Macro	100000	2	866,200.1 ns	23,908.85 ns	1,310.52 ns	-	473 B
Report	100000	2	872,946.7 ns	53,363.30 ns	2,925.02 ns	-	6545 B
Matrix	100000	10	979,970.8 ns	18,615.58 ns	1,020.38 ns	-	1249 B
MatrixWeighted	100000	10	941,622.4 ns	159,571.29 ns	8,746.64 ns	-	1249 B
AccuracyScore	100000	10	273,308.2 ns	2,631.35 ns	144.23 ns	-	-
F1Macro	100000	10	982,164.0 ns	19,871.60 ns	1,089.23 ns	-	1665 B
Report	100000	10	991,485.6 ns	61,239.42 ns	3,356.74 ns	-	15841 B
Matrix	1000000	2	8,650,392.2 ns	104,184.14 ns	5,710.68 ns	-	324 B
MatrixWeighted	1000000	2	8,292,924.8 ns	694,332.86 ns	38,058.73 ns	-	324 B
AccuracyScore	1000000	2	1,745,566.9 ns	15,737.75 ns	862.64 ns	-	-
F1Macro	1000000	2	8,488,402.0 ns	1,416,584.63 ns	77,647.78 ns	-	484 B
Report	1000000	2	8,644,758.9 ns	123,894.61 ns	6,791.08 ns	-	6572 B
Matrix	1000000	10	9,820,834.3 ns	67,309.86 ns	3,689.48 ns	-	1260 B
MatrixWeighted	1000000	10	9,257,525.8 ns	274,673.47 ns	15,055.78 ns	-	1260 B
AccuracyScore	1000000	10	2,768,566.3 ns	69,312.22 ns	3,799.24 ns	-	-
F1Macro	1000000	10	9,778,711.2 ns	1,388,825.78 ns	76,126.23 ns	-	1676 B
Report	1000000	10	9,846,626.4 ns	82,238.16 ns	4,507.75 ns	-	15892 B

Lodestar.Text.Benchmarks.MyersGateBenchmarks-report-github

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

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	Band	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
Dp	4	84.36 ns	4.417 ns	0.242 ns	1.00	-	NA
Kernel	4	82.32 ns	0.359 ns	0.020 ns	0.98	-	NA
Dp	6	117.28 ns	16.750 ns	0.918 ns	1.00	-	NA
Kernel	6	116.84 ns	1.883 ns	0.103 ns	1.00	-	NA
Dp	8	157.75 ns	12.413 ns	0.680 ns	1.00	-	NA
Kernel	8	98.64 ns	3.593 ns	0.197 ns	0.63	-	NA
Dp	10	203.13 ns	0.828 ns	0.045 ns	1.00	-	NA
Kernel	10	111.75 ns	0.960 ns	0.053 ns	0.55	-	NA
Dp	12	269.85 ns	1.480 ns	0.081 ns	1.00	-	NA
Kernel	12	119.37 ns	5.058 ns	0.277 ns	0.44	-	NA
Dp	16	428.12 ns	6.445 ns	0.353 ns	1.00	-	NA
Kernel	16	140.85 ns	0.064 ns	0.004 ns	0.33	-	NA
Dp	24	863.38 ns	5.261 ns	0.288 ns	1.00	-	NA
Kernel	24	178.74 ns	1.590 ns	0.087 ns	0.21	-	NA
Dp	32	1,480.81 ns	33.616 ns	1.843 ns	1.00	-	NA
Kernel	32	220.62 ns	41.108 ns	2.253 ns	0.15	-	NA
Dp	48	3,260.29 ns	254.775 ns	13.965 ns	1.00	-	NA
Kernel	48	307.04 ns	16.014 ns	0.878 ns	0.09	-	NA
Dp	64	5,654.41 ns	149.414 ns	8.190 ns	1.00	-	NA
Kernel	64	390.27 ns	29.095 ns	1.595 ns	0.07	-	NA
Dp	96	12,679.77 ns	143.441 ns	7.862 ns	1.00	-	NA
Kernel	96	1,097.10 ns	33.113 ns	1.815 ns	0.09	-	NA

Against rapidfuzz, in this same run

Both sides on this VM in these minutes, which is what makes the ratio readable where the absolutes are not.

• indel
• levenshtein

Lodestar.Text.Benchmarks.PersistenceBenchmarks-report-github

As of 2026-08-20, measured at commit 6cadb6a76b4e172c36be7cdb3f541d742e24373a.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
VocabTxt	4.408 ms	2.4360 ms	0.1335 ms	117.1875	109.3750	39.0625	3.62 MB
TokenizerJsonWordPiece	11.636 ms	2.3406 ms	0.1283 ms	187.5000	171.8750	46.8750	5.72 MB
TokenizerJsonUnigram	12.475 ms	0.9455 ms	0.0518 ms	93.7500	78.1250	31.2500	4.64 MB
SpieceModel	3.898 ms	4.2486 ms	0.2329 ms	109.3750	101.5625	31.2500	3.36 MB
TfidfSave	1.907 ms	0.2954 ms	0.0162 ms	31.2500	25.3906	25.3906	2.09 MB
TfidfLoad	4.436 ms	0.5713 ms	0.0313 ms	85.9375	78.1250	23.4375	2.86 MB
EmbeddingIndexSave	6.866 ms	0.8194 ms	0.0449 ms	476.5625	476.5625	476.5625	39.64 MB
EmbeddingIndexLoad	6.026 ms	1.6341 ms	0.0896 ms	531.2500	500.0000	468.7500	35.35 MB

Lodestar.Text.Benchmarks.StopWordBenchmarks-report-github

As of 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Documents	Mean	Error	StdDev	Ratio	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Count	200	7.529 ms	1.6519 ms	0.0905 ms	1.00	500.0000	234.3750	62.5000	7.92 MB	1.00
CountWithStopWords	200	6.293 ms	0.5739 ms	0.0315 ms	0.84	390.6250	187.5000	-	6.34 MB	0.80
Hashing	200	7.361 ms	0.2672 ms	0.0146 ms	0.98	507.8125	171.8750	70.3125	8.08 MB	1.02
HashingWithStopWords	200	6.429 ms	0.4058 ms	0.0222 ms	0.85	406.2500	156.2500	-	6.52 MB	0.82
Count	1000	30.631 ms	6.0696 ms	0.3327 ms	1.00	2718.7500	1031.2500	531.2500	38.64 MB	1.00
CountWithStopWords	1000	24.671 ms	0.7855 ms	0.0431 ms	0.81	1968.7500	781.2500	250.0000	30.99 MB	0.80
Hashing	1000	29.081 ms	1.9117 ms	0.1048 ms	0.95	2562.5000	750.0000	593.7500	39.55 MB	1.02
HashingWithStopWords	1000	25.044 ms	1.5254 ms	0.0836 ms	0.82	2031.2500	625.0000	250.0000	31.83 MB	0.82

Lodestar.Text.Benchmarks.VectorMathBenchmarks-report-github

As of 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Dim	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
Dot	384	47.98 ns	0.769 ns	0.042 ns	1.00	-	NA
L2Norm	384	48.57 ns	0.272 ns	0.015 ns	1.01	-	NA
Dot	768	94.69 ns	1.586 ns	0.087 ns	1.00	-	NA
L2Norm	768	92.76 ns	3.642 ns	0.200 ns	0.98	-	NA
Dot	1024	123.98 ns	3.653 ns	0.200 ns	1.00	-	NA
L2Norm	1024	122.67 ns	1.909 ns	0.105 ns	0.99	-	NA

Lodestar.Text.Benchmarks.VectorizerBenchmarks-report-github

As of 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

BenchmarkDotNet v0.14.0, Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763, 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	Documents	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
Count	200	3.243 ms	4.6509 ms	0.2549 ms	1.00	0.10	93.7500	39.0625	-	1.6 MB	1.00
Tfidf	200	3.178 ms	6.0860 ms	0.3336 ms	0.98	0.11	101.5625	39.0625	-	1.63 MB	1.02
CountBigrams	200	3.818 ms	0.1463 ms	0.0080 ms	1.18	0.08	171.8750	109.3750	-	2.78 MB	1.74
Hashing	200	2.905 ms	0.8131 ms	0.0446 ms	0.90	0.07	93.7500	23.4375	-	1.6 MB	1.00
Count	1000	7.195 ms	1.2123 ms	0.0664 ms	1.00	0.01	484.3750	343.7500	62.5000	7.83 MB	1.00
Tfidf	1000	7.478 ms	0.6631 ms	0.0363 ms	1.04	0.01	484.3750	312.5000	93.7500	7.97 MB	1.02
CountBigrams	1000	11.969 ms	1.8187 ms	0.0997 ms	1.66	0.02	906.2500	375.0000	265.6250	13.42 MB	1.71
Hashing	1000	7.065 ms	0.7149 ms	0.0392 ms	0.98	0.01	492.1875	156.2500	70.3125	7.85 MB	1.00

Against rapidfuzz, in this same run

Both sides on this VM in these minutes, which is what makes the ratio readable where the absolutes are not.

• indel
• levenshtein
• metrics
• persistence

compare-indel

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

Python: rapidfuzz 3.14.5 (py 3.12.13)
C#:     Lodestar.Text on .NET 10.0.11 (mode Utf16Unit)

length	Python ns/pair	C# ns/pair	speedup (py/C#)
8	111.1	24.1	4.60x C# faster
32	157.8	75.3	2.10x C# faster
128	478.8	957.1	2.00x Py faster
512	4873.8	11410.3	2.34x Py faster

Note: Indel is len(a)+len(b)-2*LCS on both sides, so this compares the subsequence kernels. Lodestar's is a rolling-row dynamic program (#273).

compare-levenshtein

As of 2026-08-20, measured at commit 6bde87c833c7f936cefed0881eabbccbb2cb2f1f.

Python: rapidfuzz 3.14.5 (py 3.12.13)
C#:     Lodestar.Text on .NET 10.0.11 (mode Utf16Unit)

length	Python ns/pair	C# ns/pair	speedup (py/C#)
8	151.0	18.4	8.21x C# faster
32	267.5	132.9	2.01x C# faster
128	1828.1	1435.6	1.27x C# faster
512	15596.2	16268.0	1.04x Py 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 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

Python: {'scikit-learn': '1.9.0', 'numpy': '2.5.1'} (py 3.12.13)
C#:     Lodestar on .NET 10.0.11

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
confusion_matrix_n1000_k2	0.009	0.993	105.68x	0.009	0.993	105.63x
accuracy_n1000_k2	0.001	0.534	571.54x	0.001	0.534	571.51x
precision_recall_f1_macro_n1000_k2	0.008	1.760	224.82x	0.008	1.760	224.83x
classification_report_n1000_k2	0.011	6.740	620.44x	0.011	6.739	620.32x
roc_auc_binary_n1000_k2	0.016	1.943	123.83x	0.016	1.943	123.82x
balanced_accuracy_n1000_k2	0.008	1.076	139.00x	0.008	1.076	139.02x
matthews_n1000_k2	0.007	2.016	269.46x	0.007	2.016	269.47x
cohen_kappa_n1000_k2	0.007	1.131	150.87x	0.007	1.131	150.88x
mse_n1000_k2	0.002	0.317	131.20x	0.002	0.317	131.21x
mae_n1000_k2	0.002	0.316	130.53x	0.002	0.316	130.53x
median_ae_n1000_k2	0.007	0.329	49.18x	0.007	0.329	49.19x
r2_n1000_k2	0.002	0.379	152.21x	0.002	0.379	152.19x
confusion_matrix_n1000_k10	0.009	1.012	106.63x	0.009	1.012	106.63x
accuracy_n1000_k10	0.001	0.545	583.35x	0.001	0.545	583.36x
precision_recall_f1_macro_n1000_k10	0.008	1.792	213.50x	0.008	1.788	212.97x
classification_report_n1000_k10	0.016	7.011	443.27x	0.016	7.011	443.34x
roc_auc_ovr_macro_n1000_k10	0.545	9.916	18.18x	0.545	9.915	18.18x
balanced_accuracy_n1000_k10	0.008	1.073	128.83x	0.008	1.073	128.83x
matthews_n1000_k10	0.008	2.035	253.63x	0.008	2.035	253.64x
cohen_kappa_n1000_k10	0.008	1.118	133.02x	0.008	1.118	133.03x
mse_n1000_k10	0.002	0.312	128.97x	0.002	0.312	128.97x
mae_n1000_k10	0.002	0.313	129.07x	0.002	0.313	129.06x
median_ae_n1000_k10	0.007	0.324	48.23x	0.007	0.324	48.23x
r2_n1000_k10	0.002	0.378	151.81x	0.002	0.378	151.82x
confusion_matrix_n100000_k2	1.002	10.733	10.72x	1.001	10.733	10.72x
accuracy_n100000_k2	0.164	3.759	22.92x	0.164	3.759	22.92x
precision_recall_f1_macro_n100000_k2	0.854	12.283	14.39x	0.854	12.282	14.38x
classification_report_n100000_k2	0.866	26.883	31.04x	0.866	26.880	31.04x
roc_auc_binary_n100000_k2	3.894	26.965	6.92x	3.894	26.964	6.92x
balanced_accuracy_n100000_k2	0.858	10.781	12.57x	0.858	10.781	12.57x
matthews_n100000_k2	0.838	21.553	25.72x	0.838	21.552	25.72x
cohen_kappa_n100000_k2	0.839	10.846	12.92x	0.839	10.845	12.92x
mse_n100000_k2	0.242	0.467	1.93x	0.242	0.467	1.93x
mae_n100000_k2	0.238	0.458	1.93x	0.238	0.458	1.93x
median_ae_n100000_k2	0.784	1.797	2.29x	0.806	1.797	2.23x
r2_n100000_k2	0.235	0.711	3.02x	0.235	0.711	3.02x
confusion_matrix_n100000_k10	0.941	10.746	11.42x	0.941	10.743	11.42x
accuracy_n100000_k10	0.276	3.767	13.67x	0.276	3.767	13.67x
precision_recall_f1_macro_n100000_k10	0.957	12.957	13.54x	0.957	12.956	13.54x
classification_report_n100000_k10	0.970	29.565	30.49x	0.970	29.563	30.49x
roc_auc_ovr_macro_n100000_k10	36.458	219.479	6.02x	36.457	219.460	6.02x
balanced_accuracy_n100000_k10	0.968	10.847	11.20x	0.968	10.847	11.20x
matthews_n100000_k10	0.942	22.494	23.88x	0.942	22.493	23.88x
cohen_kappa_n100000_k10	1.160	10.869	9.37x	1.160	10.868	9.37x
mse_n100000_k10	0.242	0.468	1.93x	0.242	0.468	1.93x
mae_n100000_k10	0.238	0.455	1.91x	0.238	0.455	1.91x
median_ae_n100000_k10	0.810	1.800	2.22x	0.864	1.800	2.08x
r2_n100000_k10	0.236	0.662	2.81x	0.236	0.662	2.81x
confusion_matrix_n1000000_k2	8.454	101.835	12.05x	8.454	101.810	12.04x
accuracy_n1000000_k2	1.829	33.808	18.48x	1.829	33.803	18.48x
precision_recall_f1_macro_n1000000_k2	8.629	109.125	12.65x	8.628	109.108	12.65x
classification_report_n1000000_k2	8.646	213.302	24.67x	8.645	213.281	24.67x
roc_auc_binary_n1000000_k2	55.291	294.713	5.33x	55.285	294.655	5.33x
balanced_accuracy_n1000000_k2	8.706	101.199	11.62x	8.704	101.180	11.62x
matthews_n1000000_k2	8.453	202.799	23.99x	8.451	202.776	23.99x
cohen_kappa_n1000000_k2	8.460	101.376	11.98x	8.459	101.363	11.98x
mse_n1000000_k2	2.420	2.572	1.06x	2.420	2.572	1.06x
mae_n1000000_k2	2.400	2.583	1.08x	2.400	2.583	1.08x
median_ae_n1000000_k2	7.199	14.912	2.07x	7.270	14.909	2.05x
r2_n1000000_k2	2.423	5.116	2.11x	2.423	5.115	2.11x
confusion_matrix_n1000000_k10	9.449	101.779	10.77x	9.448	101.735	10.77x
accuracy_n1000000_k10	2.755	34.039	12.36x	2.755	34.038	12.36x
precision_recall_f1_macro_n1000000_k10	9.639	116.969	12.13x	9.639	116.957	12.13x
classification_report_n1000000_k10	9.714	241.324	24.84x	9.713	241.292	24.84x
balanced_accuracy_n1000000_k10	9.631	102.976	10.69x	9.630	102.962	10.69x
matthews_n1000000_k10	9.455	213.243	22.55x	9.454	213.230	22.55x
cohen_kappa_n1000000_k10	9.462	102.936	10.88x	9.461	102.921	10.88x
mse_n1000000_k10	2.406	2.848	1.18x	2.405	2.847	1.18x
mae_n1000000_k10	2.381	2.877	1.21x	2.381	2.876	1.21x
median_ae_n1000000_k10	7.033	15.125	2.15x	7.090	15.123	2.13x
r2_n1000000_k10	2.849	5.432	1.91x	2.849	5.431	1.91x

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-persistence

As of 2026-08-20, measured at commit abfaf002802c5e98262d1a1d5072a1f6bfa644b0.

Python: {'tokenizers': '0.23.1', 'sentencepiece': '0.2.2', 'scikit-learn': '1.9.0', 'numpy': '2.5.1'} (py 3.12.13)
C#:     Lodestar on .NET 10.0.11

operation	C# ms	Py ms	wall	C# cpu	Py cpu	cpu
vocab_txt	4.424	10.675	2.41x	4.721	10.674	2.26x
tokenizer_json_wordpiece	12.140	17.650	1.45x	12.495	17.648	1.41x
tokenizer_json_unigram	12.702	41.925	3.30x	13.247	41.923	3.16x
spiece_model	4.660	29.713	6.38x	4.911	29.710	6.05x
tfidf_save	1.999	2.483	1.24x	2.139	2.482	1.16x
tfidf_load	4.211	4.444	1.06x	4.454	4.443	1.00x
embedding_index_save	7.216	4.686	0.65x	7.814	4.685	0.60x
embedding_index_load	6.634	1.693	0.26x	7.589	1.693	0.22x

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.