[Perf] Linux/x64: 76 Regressions on 6/24/2026 7:31:32 AM +00:00

[performanceautofiler]

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	0d1a20fb14109f277df06ebee3f83c964f9dcc61
Compare	e3cbc00cd3e0b511ad36dc240010eb2b9eefc3b7
Diff	Diff
Configs	AOT:true, CompilationMode:wasm, RunKind:micro

Regressions in System.Text.Json.Tests.Perf_Reader

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.51 μs	14.64 μs	1.95	0.07	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.47 μs	21.22 μs	1.85	0.07	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.37 μs	13.25 μs	1.80	0.10	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.89 μs	3.37 μs	1.78	0.10	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.74 μs	20.70 μs	1.76	0.05	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	74.89 μs	131.49 μs	1.76	0.10	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	74.61 μs	129.53 μs	1.74	0.08	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	937.62 ns	1614.91 ns	1.72	0.08	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.01 μs	3.46 μs	1.72	0.10	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.52 μs	22.87 μs	1.69	0.08	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	3.04 μs	5.09 μs	1.67	0.06	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.29 μs	3.83 μs	1.67	0.09	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	985.18 ns	1636.79 ns	1.66	0.07	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	88.60 μs	146.70 μs	1.66	0.08	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	87.39 μs	143.83 μs	1.65	0.13	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	3.08 μs	5.02 μs	1.63	0.07	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.90 μs	14.45 μs	1.62	0.11	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.93 μs	14.44 μs	1.62	0.06	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.45 μs	23.10 μs	1.60	0.06	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	3.53 μs	5.62 μs	1.59	0.11	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.12 μs	1.74 μs	1.55	0.10	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.31 μs	2.02 μs	1.54	0.07	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	131.00 ns	201.57 ns	1.54	0.16	False
ReadSpanEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	6.81 μs	10.43 μs	1.53	0.05	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.49 μs	3.78 μs	1.52	0.09	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	6.88 μs	10.42 μs	1.51	0.08	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	4.19 μs	6.30 μs	1.51	0.14	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	162.69 ns	242.26 ns	1.49	0.17	False
ReadReturnBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.80 μs	11.01 μs	1.41	0.08	False
ReadSingleSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	175.79 ns	247.00 ns	1.41	0.14	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.97 μs	11.16 μs	1.40	0.04	False
ReadMultiSpanSequenceEmptyLoop - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	330.62 ns	410.23 ns	1.24	0.14	False

Test Report

Repro

General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Text.Json.Tests.Perf_Reader*'

Details

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadReturnBytes(IsDataCompact: False, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Reader.ReadMultiSpanSequenceEmptyLoop(IsDataCompact: False, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

[performanceautofiler]

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	0d1a20fb14109f277df06ebee3f83c964f9dcc61
Compare	e3cbc00cd3e0b511ad36dc240010eb2b9eefc3b7
Diff	Diff
Configs	AOT:true, CompilationMode:wasm, RunKind:micro

Regressions in System.Text.Json.Tests.Perf_Segment

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ReadSingleSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.42 μs	13.89 μs	1.87	0.10	False
ReadSingleSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	736.21 μs	1348.68 μs	1.83	0.08	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	752.67 μs	1374.41 μs	1.83	0.09	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.45 μs	13.47 μs	1.81	0.07	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.46 μs	13.46 μs	1.80	0.13	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.42 μs	13.14 μs	1.77	0.10	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	754.95 μs	1335.41 μs	1.77	0.14	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	77.07 μs	136.00 μs	1.76	0.13	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	765.96 μs	1345.82 μs	1.76	0.10	False
ReadSingleSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	74.85 μs	131.05 μs	1.75	0.12	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	74.97 μs	131.24 μs	1.75	0.06	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.78 μs	13.47 μs	1.73	0.08	False
ReadSingleSegmentSequenceByN - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	76.28 μs	131.49 μs	1.72	0.08	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	77.03 μs	132.28 μs	1.72	0.03	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	796.67 μs	1345.74 μs	1.69	0.08	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	884.07 μs	1431.06 μs	1.62	0.07	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.88 μs	14.35 μs	1.62	0.07	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	87.82 μs	141.23 μs	1.61	0.06	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	88.82 μs	141.64 μs	1.59	0.10	False
ReadMultiSegmentSequence - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	888.78 μs	1406.44 μs	1.58	0.10	False
ReadMultiSegmentSequenceUsingSpan - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	10.40 μs	13.42 μs	1.29	0.09	False

Test Report

Repro

General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Text.Json.Tests.Perf_Segment*'

Details

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequence(TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequence(TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 4096, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 8192, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 4096, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 8192, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 8192, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 8192, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 4096, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequence(TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 8192, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 4096, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 4096, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 4096, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 8192, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 4096, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 4096, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 8192, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 4096, TestCase: Json40KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequence(segmentSize: 8192, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Tests.Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 8192, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

---

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	0d1a20fb14109f277df06ebee3f83c964f9dcc61
Compare	e3cbc00cd3e0b511ad36dc240010eb2b9eefc3b7
Diff	Diff
Configs	AOT:true, CompilationMode:wasm, RunKind:micro

Regressions in System.Text.Json.Document.Tests.Perf_DocumentParse

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.21 μs	1.94 μs	1.60	0.12	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	952.99 μs	1510.27 μs	1.58	0.09	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.38 μs	2.05 μs	1.49	0.09	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	312.15 ns	402.94 ns	1.29	0.14	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	3.60 μs	4.49 μs	1.25	0.15	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	679.65 ns	765.76 ns	1.13	0.23	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	6.02 μs	6.65 μs	1.10	0.08	False
Parse - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.98 ms	9.63 ms	1.07	0.04	False

Test Report

Repro

General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Text.Json.Document.Tests.Perf_DocumentParse*'

Details

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: False, TestCase: BasicJson)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: False, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: False, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: False, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: True, TestCase: BasicJson)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: True, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: True, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: True, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

[performanceautofiler]

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	0d1a20fb14109f277df06ebee3f83c964f9dcc61
Compare	e3cbc00cd3e0b511ad36dc240010eb2b9eefc3b7
Diff	Diff
Configs	AOT:true, CompilationMode:wasm, RunKind:micro

Regressions in System.Text.Json.Node.Tests.Perf_ParseThenWrite

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	86.33 μs	97.74 μs	1.13	0.10	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.36 μs	15.08 μs	1.13	0.08	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	55.89 μs	62.33 μs	1.12	0.02	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	5.78 ms	6.40 ms	1.11	0.06	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	20.00 μs	21.72 μs	1.09	0.05	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	49.68 μs	53.87 μs	1.08	0.04	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.31 μs	1.40 μs	1.07	0.07	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.70 μs	8.19 μs	1.06	0.06	False

Test Report

Repro

General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Text.Json.Node.Tests.Perf_ParseThenWrite*'

Details

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: HelloWorld)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Node.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

---

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	0d1a20fb14109f277df06ebee3f83c964f9dcc61
Compare	e3cbc00cd3e0b511ad36dc240010eb2b9eefc3b7
Diff	Diff
Configs	AOT:true, CompilationMode:wasm, RunKind:micro

Regressions in System.Text.Json.Document.Tests.Perf_ParseThenWrite

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	35.87 μs	44.83 μs	1.25	0.05	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	19.97 μs	23.80 μs	1.19	0.04	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	26.71 μs	31.72 μs	1.19	0.05	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.77 μs	9.15 μs	1.18	0.09	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.61 ms	3.05 ms	1.17	0.05	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	3.74 μs	4.36 μs	1.16	0.06	False
ParseThenWrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	9.46 μs	10.85 μs	1.15	0.08	False

Test Report

Repro

General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Text.Json.Document.Tests.Perf_ParseThenWrite*'

Details

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: BroadTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: DeepTree)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json4KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: LotsOfStrings)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json400KB)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: Json400B)

ETL Files

Histogram

JIT Disasms

System.Text.Json.Document.Tests.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: LotsOfNumbers)

ETL Files

Histogram

JIT Disasms

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

[LoopedBard3]

🔍 Automated Triage Analysis

Summary: Massive regression across 74 System.Text.Json reader/parser tests (+6–95%, median +60%) on WASM AOT, attributed to commit 65b2508c9e09 which introduced SearchValues-based vectorized whitespace skipping in Utf8JsonReader.SkipWhiteSpace. Finding confidence 4/5 (excellent historical data, 295+ runs per test, CV < 3%), regression confidence 4/5 (direct code path hit, surgical change, step change pattern across all affected tests). Bisection skipped — autofile run configuration is not supported by the bisection harness.

Finding Confidence: 4/5 (analysis accuracy)
Regression Confidence: 4/5 (likelihood of true regression)

🤖 Proposed Next Actions: Uncertain — flag for re-triage — 1 done, 1 dry-run (rule R5_look_again — see Full Analysis for the table). · 🦆 Re-analysis: ⚠️ disputes (overconfident, weak-bisection-evidence, culprit-unsupported-by-bisection)

📊 Full Analysis (click to expand)

Summary: Massive regression across 74 System.Text.Json reader/parser tests (+6–95%, median +60%) on WASM AOT, attributed to commit 65b2508c9e09 which introduced SearchValues-based vectorized whitespace skipping in Utf8JsonReader.SkipWhiteSpace. Finding confidence 4/5 (excellent historical data, 295+ runs per test, CV < 3%), regression confidence 4/5 (direct code path hit, surgical change, step change pattern across all affected tests).

Issue Overview

• Issue: [Perf] Linux/x64: 76 Regressions on 6/24/2026 7:31:32 AM +00:00 #75553 - [Perf] Linux/x64: 76 Regressions on 6/24/2026 7:31:32 AM +00:00
• Date: 2026-06-24 | Queue: Ubuntu.2204.Amd64.Viper.Perf | OS/Arch: linux/x64 (WASM AOT)
• Commits: 0d1a20fb → e3cbc00c (16 commits)
• Labels: untriaged, perf-regression, arch-x64, arch-wasm, wasm-aot, os-linux, kind-micro, aot-true, compilationmode-wasm
• Runtime Stack: WASM + Mono
• Historical Data: Available (295-299 runs per test, 1237 total runs examined)

Test Analysis

Test Name	Baseline	Compare	Δ%	μ	σ	n	z	Thresh%	Noise?	Assessment
Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, Test	7514.0	14643.0	+94.9%	7483.1	150.2	295	47.7	5%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequence(TestCase: Json4KB)	7420.0	13894.0	+87.2%	7504.1	214.6	295	29.8	5.72%	N	Step change, stable
Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: BroadTre	11469.0	21220.0	+85.0%	11474.1	195.6	297	49.8	5%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequence(TestCase: Json400KB)	736207.0	1348683.0	+83.2%	748555.0	19533.2	297	30.7	5.22%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 4096, TestCas	752671.0	1374406.0	+82.6%	760530.1	25594.1	297	24.0	6.73%	N	Step change, stable
Perf_Segment.ReadMultiSegmentSequence(segmentSize: 8192, TestCase: Jso	7451.0	13472.0	+80.8%	7533.7	140.1	299	42.4	5%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 4096, TestCas	7460.0	13458.0	+80.4%	7570.5	187.2	295	31.5	5%	N	Step change, stable
Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: Json4KB)	7369.0	13254.0	+79.9%	7437.4	178.1	297	32.7	5%	N	Step change, stable
Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: LotsOfSt	1887.0	3367.0	+78.4%	1925.0	52.4	297	27.5	5.45%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 8192, TestCas	7416.0	13141.0	+77.2%	7454.5	177.0	298	32.1	5%	N	Step change, stable
Perf_Segment.ReadSingleSegmentSequenceByN(numberOfBytes: 8192, TestCas	754951.0	1335412.0	+76.9%	759899.7	29370.7	296	19.6	7.73%	N	Step change, stable
Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 8192, Test	77070.0	135999.0	+76.5%	77264.1	4644.6	296	12.6	12.02%	N	Step change, stable
Perf_Reader.ReadSingleSpanSequenceEmptyLoop(IsDataCompact: False, Test	11737.0	20704.0	+76.4%	11534.6	237.3	296	38.6	5%	N	Step change, stable
Perf_Segment.ReadMultiSegmentSequenceUsingSpan(segmentSize: 4096, Test	765962.0	1345824.0	+75.7%	778129.5	32206.2	296	17.6	8.28%	N	Step change, stable
Perf_Reader.ReadSpanEmptyLoop(IsDataCompact: False, TestCase: Json40KB	74886.0	131489.0	+75.6%	74332.4	3189.0	296	17.9	8.58%	N	Step change, stable
Doc.Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: T	6023.0	6655.0	+10.5%	6049.1	440.3	299	1.4	14.56%	Y	Within variance
Node.Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase	1306.0	1400.0	+7.2%	1302.2	65.7	296	1.5	10.09%	Y	Within variance

Legend: μ=mean (μs), σ=stddev, n=runs, z=z-score, Thresh%=adaptive threshold. Values are in microseconds.

(Plus 59 additional tests with similar regression patterns across Perf_Reader, Perf_Segment, Perf_DocumentParse, and Perf_ParseThenWrite families)

Finding Confidence: 4/5 - Excellent historical data (295+ runs), low variance (CV 1.7–6.4%), no data quality issues. Not capped because historical baseline is comprehensive.

Candidate Stack Relevance

Commit	Touched Paths (summary)	Relevance	Notes
65b2508c9e09	src/libraries/System.Text.Json/src/.../Reader/	Shared	BCL change to Utf8JsonReader; uses SearchValues/IndexOfAnyExcept which may not vectorize efficiently on WASM
a79c6fbd8514	src/mono/wasm/, src/mono/mono/	Relevant	Adds Math intrinsics for wasm; no JSON code path impact
feb1fe23e41b	src/coreclr/jit/codegenwasm.cpp, src/coreclr/jit/fgwasm.*	LessLikely	CoreCLR JIT wasm EH fixes; not exercised by Mono WASM AOT runs
43b739fcc63f	src/mono/ AOT marshalling	Relevant	Mono AOT wrapper compilation; no JSON reader path impact
b06f62fc8e16	src/mono/ AOT jmp opcode	Relevant	Mono AOT method compilation; unlikely to affect hot loops

Confirmed Regressions

System.Text.Json Utf8JsonReader Family (32 tests, +50–95%)

• Delta: Median +67% across ReadSpanEmptyLoop, ReadSingleSpanSequenceEmptyLoop, ReadMultiSpanSequenceEmptyLoop, ReadReturnBytes families
• Historical Context: μ stable with CV 1.7–2.7%; z-scores 27–50; adaptive threshold 5%
• Assessment: Extreme step change. Complete distributional separation between before and after series. No overlap whatsoever.
• Finding Confidence: 4/5 - Exceptional historical data quality, 295+ runs, low CV, clear step change
• Regression Confidence: 4/5 - Direct code path hit (SkipWhiteSpace is the hot loop for whitespace-heavy JSON), surgical change, consistent pattern across all reader variants
• Likely Cause: 65b2508c9e09 - "Use SearchValues in Utf8JsonReader.SkipWhiteSpace (#129701)"
  • Affected Areas: System.Text.Json / Utf8JsonReader / BCL SearchValues
  • Stack Relevance: Shared (BCL library change affects all runtimes including WASM)
  • Rationale: This commit replaces the scalar whitespace-skipping loop in Utf8JsonReader.SkipWhiteSpace with SearchValues<byte>.IndexOfAnyExcept(). On native x64, this uses SIMD vectorization for ~2× speedup on indented JSON. However, on WASM AOT (Mono), SearchValues/IndexOfAnyExcept likely cannot use hardware SIMD (or the WASM SIMD implementation has significant overhead), resulting in the vectorized path being ~2× slower than the original scalar loop. All affected tests use IsDataCompact: False (indented/pretty-printed JSON), confirming the whitespace-skipping path is the bottleneck. The #if NET guard means WASM AOT also takes this path.
  • Candidates Considered:
    • 65b2508c9e09 - Direct Utf8JsonReader.SkipWhiteSpace modification; timing aligns exactly
    • a79c6fbd8514 - WASM math intrinsics; no JSON path impact
    • feb1fe23e41b - CoreCLR JIT wasm EH; not exercised by Mono WASM
• Verification Status: Not verified. Recommend RunWithBuildAtHash on 65b2508c9e09 vs parent 5bd107431d14.

System.Text.Json Perf_Segment Family (22 tests, +29–87%)

• Delta: Median +75% across ReadSingleSegmentSequence, ReadMultiSegmentSequence, ReadMultiSegmentSequenceUsingSpan, ReadSingleSegmentSequenceByN
• Historical Context: z-scores 12–50; CV 2–6%
• Assessment: Same step change pattern. Segment tests use multi-segment reader which delegates to SkipWhiteSpaceMultiSegment → SkipWhiteSpace.
• Finding Confidence: 4/5 - Same excellent data quality
• Regression Confidence: 4/5 - Same root cause via shared SkipWhiteSpace path
• Likely Cause: Same as above - 65b2508c9e09

System.Text.Json DocumentParse & ParseThenWrite (20 tests, +6–60%)

• Delta: Median +25% for DocumentParse, +13% for ParseThenWrite
• Historical Context: z-scores 2.5–33; lower deltas reflect less whitespace-dominated workload
• Assessment: Step change consistent with same root cause but attenuated because these tests spend less time in whitespace skipping relative to parsing/writing.
• Finding Confidence: 4/5
• Regression Confidence: 4/5 - Same pattern, same timing
• Likely Cause: Same - 65b2508c9e09

Noise Tests (2 tests)

• Perf_DocumentParse.Parse(IsDataIndented: True, TestRandomAccess: True, TestCase: Json400B): +10.5%, z=1.4 — high historical variance from outliers; within adaptive threshold of 14.6%
• Perf_ParseThenWrite.ParseThenWrite(IsDataIndented: True, TestCase: HelloWorld): +7.2%, z=1.5 — high CV (5.0%) from outliers; within adaptive threshold of 10.1%

Open Questions

• Verification Needed: RunWithBuildAtHash on commit 65b2508c9e09 vs its parent (5bd107431d14) would confirm the regression is isolated to this single commit
• WASM SearchValues Performance: Need to understand why SearchValues<byte>.IndexOfAnyExcept is slower on WASM AOT than a scalar loop — possible causes include: WASM SIMD overhead in Mono AOT, lack of vectorization, or SearchValues initialization cost per call
• Bisection Status: Not performed — the commit attribution is high-confidence based on code path analysis, but verification would elevate confidence to 5/5

Recommended Actions

1. Verify regression: Run bisection with RunWithBuildAtHash on 65b2508c9e09 vs 5bd107431d14 to confirm this single commit is the cause
2. Fix: Guard the vectorized SkipWhiteSpace path with a WASM-specific check (e.g., #if NET && !BROWSER) or ensure SearchValues/IndexOfAnyExcept falls back to an efficient scalar implementation on WASM AOT
3. Investigate SearchValues on WASM: Profile the IndexOfAnyExcept implementation under Mono WASM AOT to understand whether the regression is in SearchValues initialization, the SIMD path, or call overhead
4. Contact PR author: Notify the author of PR #129701 about the WASM regression

🤖 Proposed Next Actions

Rule: R5_look_again — Uncertain — flag for re-triage.

Confidence is sufficient (finding 4/5, regression 4/5), but bisection was skipped because the autofile run configuration is not supported by the bisection harness.

#	Action	Status	Details
1	Add label	✅ Done	Look Again
2	Remove label	⏭️ Skipped (dry-run)	untriaged

🦆 Re-analysis (advisory)

⚠️ Disputes the analysis — advisory only, not applied

Flags: overconfident, weak-bisection-evidence, culprit-unsupported-by-bisection

Rationale: The bisection was skipped (status=skipped, culprit=none), so there is no convergence evidence supporting the 65b2508c9e09 commit or the 4/5 change confidence. The consistency across 10 related System.Text.Json benchmarks is suggestive but not a substitute for bisection confirmation. Lowering change confidence to 3/5 to reflect that the culprit is inferred from timing correlation alone, not validated by bisection.

Suggested confidence: Change Confidence → 3/5 (advisory — not applied)

This critique is advisory in the current configuration and did not change the confidences, culprit, or proposed actions above.

---

Generated by PerfTriageAgent on 2026-07-02 13:13 UTC