[Perf] Linux/arm64: 8 Regressions on 10/7/2023 2:03:18 PM #93401

performanceautofiler · 2023-10-12T07:49:32Z

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.IterateFor<String>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ReadOnlySpan - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	205.40 ns	222.33 ns	1.08	0.03	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.Collections.IterateFor&lt;String&gt;*'

Payloads

Baseline
Compare

System.Collections.IterateFor<String>.ReadOnlySpan(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 222.32651915631124 > 215.6400041349606.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 9/23/2023 7:05:34 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -372.39294984996354 (T) = (0 -222.2885707448892) / Math.Sqrt((0.004433913439093746 / (12)) + (0.010203089316673741 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.08250910009314494 = (205.3456832148222 - 222.2885707448892) / 205.3456832148222 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Text.Perf_Ascii

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ToUpper_Bytes_Chars - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	23.88 ns	28.56 ns	1.20	0.14	False
ToLower_Bytes_Chars - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	23.94 ns	28.02 ns	1.17	0.15	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.Perf_Ascii*'

Payloads

Baseline
Compare

System.Text.Perf_Ascii.ToUpper_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 28.559514717492952 > 24.79856478599641.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -20.76254857027171 (T) = (0 -28.104504760503044) / Math.Sqrt((0.3645918407156724 / (12)) + (0.05786128073273915 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1734346247423125 = (23.95063531270421 - 28.104504760503044) / 23.95063531270421 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

System.Text.Perf_Ascii.ToLower_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 28.019010068564185 > 24.869754563726246.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -20.25521761383546 (T) = (0 -28.24787440830941) / Math.Sqrt((0.2906103764827192 / (12)) + (0.13260424713375574 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.18248282820609205 = (23.88861278532339 - 28.24787440830941) / 23.88861278532339 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Buffers.Text.Tests.Base64Tests

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ConvertToBase64CharArray - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	427.30 ns	530.37 ns	1.24	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.Buffers.Text.Tests.Base64Tests*'

Payloads

Baseline
Compare

System.Buffers.Text.Tests.Base64Tests.ConvertToBase64CharArray(NumberOfBytes: 1000)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 530.3676104438226 > 447.5848694018119.
IsChangePoint: Marked as a change because one of 4/12/2023 8:39:35 PM, 4/28/2023 1:33:49 PM, 5/3/2023 7:05:50 AM, 7/1/2023 8:07:46 AM, 7/17/2023 4:07:27 PM, 8/4/2023 10:42:53 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -51.53911793942635 (T) = (0 -525.548318232403) / Math.Sqrt((21.142287272695157 / (12)) + (13.333299949151087 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.24338488874057676 = (422.6754909050972 - 525.548318232403) / 422.6754909050972 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.CreateAddAndClear<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
List - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	1.50 μs	1.80 μs	1.20	0.24	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.Collections.CreateAddAndClear&lt;Int32&gt;*'

Payloads

Baseline
Compare

System.Collections.CreateAddAndClear<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 1.7977876246594005 > 1.5936733124485132.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -10.988586748442112 (T) = (0 -1838.0158013387618) / Math.Sqrt((3446.231724137695 / (12)) + (2498.4499069713015 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1884698700010806 = (1546.5396706582815 - 1838.0158013387618) / 1546.5396706582815 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.Perf_Frozen<NotKnownComparable>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Contains_True - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	13.49 ns	15.78 ns	1.17	0.02	True

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.Collections.Perf_Frozen&lt;NotKnownComparable&gt;*'

Payloads

Baseline
Compare

System.Collections.Perf_Frozen<NotKnownComparable>.Contains_True(Count: 4)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 15.776818896170987 > 14.160815293970957.
IsChangePoint: Marked as a change because one of 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -59.943852744772315 (T) = (0 -15.728146954519806) / Math.Sqrt((0.002666869030084915 / (12)) + (0.00715756535636858 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1673708091342549 = (13.473137097015565 - 15.728146954519806) / 13.473137097015565 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Memory.Span<Char>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
IndexOfAnyThreeValues - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	44.00 ns	49.23 ns	1.12	0.10	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.Memory.Span&lt;Char&gt;*'

Payloads

Baseline
Compare

System.Memory.Span<Char>.IndexOfAnyThreeValues(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 49.22978510694759 > 46.15877453468914.
IsChangePoint: Marked as a change because one of 4/28/2023 1:33:49 PM, 5/3/2023 7:05:50 AM, 7/1/2023 8:07:46 AM, 7/17/2023 4:07:27 PM, 8/5/2023 6:15:42 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -14.473174356749167 (T) = (0 -49.362602321879855) / Math.Sqrt((0.10814897384925021 / (12)) + (0.7497474708297018 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.12021896215335026 = (44.06513725405271 - 49.362602321879855) / 44.06513725405271 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.AddGivenSize<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
List - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	1.44 μs	1.66 μs	1.15	0.35	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.Collections.AddGivenSize&lt;Int32&gt;*'

Payloads

Baseline
Compare

System.Collections.AddGivenSize<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 1.66042285032892 > 1.4533978942651908.
IsChangePoint: Marked as a change because one of 7/18/2023 9:59:03 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -15.426483125143072 (T) = (0 -1658.5832139364875) / Math.Sqrt((4077.5324724249035 / (12)) + (116.7192038612997 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.21401884300283375 = (1366.192315297215 - 1658.5832139364875) / 1366.192315297215 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

The text was updated successfully, but these errors were encountered:

EgorBo · 2023-10-12T16:32:08Z

Either the JIT change #92768 cc @jakobbotsch or BCL changes (#93125 or #93126)

ghost · 2023-10-12T16:32:26Z

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

Issue Details

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.IterateFor<String>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ReadOnlySpan - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	205.40 ns	222.33 ns	1.08	0.03	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.Collections.IterateFor&lt;String&gt;*'

Payloads

Baseline
Compare

System.Collections.IterateFor<String>.ReadOnlySpan(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 222.32651915631124 > 215.6400041349606.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 9/23/2023 7:05:34 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -372.39294984996354 (T) = (0 -222.2885707448892) / Math.Sqrt((0.004433913439093746 / (12)) + (0.010203089316673741 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.08250910009314494 = (205.3456832148222 - 222.2885707448892) / 205.3456832148222 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Text.Perf_Ascii

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ToUpper_Bytes_Chars - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	23.88 ns	28.56 ns	1.20	0.14	False
ToLower_Bytes_Chars - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	23.94 ns	28.02 ns	1.17	0.15	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.Perf_Ascii*'

Payloads

Baseline
Compare

System.Text.Perf_Ascii.ToUpper_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 28.559514717492952 > 24.79856478599641.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -20.76254857027171 (T) = (0 -28.104504760503044) / Math.Sqrt((0.3645918407156724 / (12)) + (0.05786128073273915 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1734346247423125 = (23.95063531270421 - 28.104504760503044) / 23.95063531270421 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

System.Text.Perf_Ascii.ToLower_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 28.019010068564185 > 24.869754563726246.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -20.25521761383546 (T) = (0 -28.24787440830941) / Math.Sqrt((0.2906103764827192 / (12)) + (0.13260424713375574 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.18248282820609205 = (23.88861278532339 - 28.24787440830941) / 23.88861278532339 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Buffers.Text.Tests.Base64Tests

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ConvertToBase64CharArray - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	427.30 ns	530.37 ns	1.24	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.Buffers.Text.Tests.Base64Tests*'

Payloads

Baseline
Compare

System.Buffers.Text.Tests.Base64Tests.ConvertToBase64CharArray(NumberOfBytes: 1000)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 530.3676104438226 > 447.5848694018119.
IsChangePoint: Marked as a change because one of 4/12/2023 8:39:35 PM, 4/28/2023 1:33:49 PM, 5/3/2023 7:05:50 AM, 7/1/2023 8:07:46 AM, 7/17/2023 4:07:27 PM, 8/4/2023 10:42:53 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -51.53911793942635 (T) = (0 -525.548318232403) / Math.Sqrt((21.142287272695157 / (12)) + (13.333299949151087 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.24338488874057676 = (422.6754909050972 - 525.548318232403) / 422.6754909050972 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.CreateAddAndClear<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
List - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	1.50 μs	1.80 μs	1.20	0.24	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.Collections.CreateAddAndClear&lt;Int32&gt;*'

Payloads

Baseline
Compare

System.Collections.CreateAddAndClear<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 1.7977876246594005 > 1.5936733124485132.
IsChangePoint: Marked as a change because one of 5/19/2023 4:22:12 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -10.988586748442112 (T) = (0 -1838.0158013387618) / Math.Sqrt((3446.231724137695 / (12)) + (2498.4499069713015 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1884698700010806 = (1546.5396706582815 - 1838.0158013387618) / 1546.5396706582815 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.Perf_Frozen<NotKnownComparable>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Contains_True - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	13.49 ns	15.78 ns	1.17	0.02	True

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.Collections.Perf_Frozen&lt;NotKnownComparable&gt;*'

Payloads

Baseline
Compare

System.Collections.Perf_Frozen<NotKnownComparable>.Contains_True(Count: 4)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 15.776818896170987 > 14.160815293970957.
IsChangePoint: Marked as a change because one of 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -59.943852744772315 (T) = (0 -15.728146954519806) / Math.Sqrt((0.002666869030084915 / (12)) + (0.00715756535636858 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.1673708091342549 = (13.473137097015565 - 15.728146954519806) / 13.473137097015565 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Memory.Span<Char>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
IndexOfAnyThreeValues - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	44.00 ns	49.23 ns	1.12	0.10	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.Memory.Span&lt;Char&gt;*'

Payloads

Baseline
Compare

System.Memory.Span<Char>.IndexOfAnyThreeValues(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 49.22978510694759 > 46.15877453468914.
IsChangePoint: Marked as a change because one of 4/28/2023 1:33:49 PM, 5/3/2023 7:05:50 AM, 7/1/2023 8:07:46 AM, 7/17/2023 4:07:27 PM, 8/5/2023 6:15:42 AM, 8/22/2023 6:29:15 PM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -14.473174356749167 (T) = (0 -49.362602321879855) / Math.Sqrt((0.10814897384925021 / (12)) + (0.7497474708297018 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.12021896215335026 = (44.06513725405271 - 49.362602321879855) / 44.06513725405271 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Run Information

Name	Value
Architecture	arm64
OS	ubuntu 20.04
Queue	AmpereUbuntu
Baseline	f55549222e4dc1469190caa56e03361ee46208ca
Compare	551a4db7917df8c321867d47ce028c726106a3c2
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.Collections.AddGivenSize<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
List - Duration of single invocation 📝 - Benchmark Source 📈 - ADX Test Multi Config Graph	1.44 μs	1.66 μs	1.15	0.35	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.Collections.AddGivenSize&lt;Int32&gt;*'

Payloads

Baseline
Compare

System.Collections.AddGivenSize<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 1.66042285032892 > 1.4533978942651908.
IsChangePoint: Marked as a change because one of 7/18/2023 9:59:03 AM, 10/6/2023 9:14:38 AM, 10/11/2023 11:44:03 AM falls between 10/2/2023 7:07:56 PM and 10/11/2023 11:44:03 AM.
IsRegressionStdDev: Marked as regression because -15.426483125143072 (T) = (0 -1658.5832139364875) / Math.Sqrt((4077.5324724249035 / (12)) + (116.7192038612997 / (6))) is less than -2.1199052992212764 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (12) + (6) - 2, .025) and -0.21401884300283375 = (1366.192315297215 - 1658.5832139364875) / 1366.192315297215 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

JIT Disasms

Docs

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

Author:	performanceautofiler[bot]
Assignees:	-
Labels:	`arch-arm64`, `os-linux`, `area-CodeGen-coreclr`, `untriaged`, `runtime-coreclr`
Milestone:	-

EgorBo · 2023-10-12T16:35:16Z

Same on Windows-Ampere-arm64: dotnet/perf-autofiling-issues#23009

EgorBo · 2023-10-12T16:42:21Z

Improvements in that commit range:

[Perf] Linux/arm64: 7 Improvements on 10/7/2023 2:03:18 PM perf-autofiling-issues#23004
[Perf] Linux/arm64: 2 Improvements on 10/9/2023 8:09:24 PM perf-autofiling-issues#23005
[Perf] Windows/arm64: 6 Improvements on 10/6/2023 11:50:29 PM perf-autofiling-issues#23011

jakobbotsch · 2023-11-06T10:48:23Z

A lot of these look bimodal/noisy, but some of them I should look closer at:

jakobbotsch · 2024-07-29T16:16:01Z

The PerfLabTests.LowLevelPerf regressed on windows-ampere only, and we stopped collecting those due to noise, so will ignore them.

Looking at the System.Collections ones instead now.

jakobbotsch · 2024-07-29T16:18:10Z

@EgorBot -arm64 -perf -commit 15af764 vs previous

// Licensed to the .NET Foundation under one or more agreements.
// The .NET Foundation licenses this file to you under the MIT license.
// See the LICENSE file in the project root for more information.

using BenchmarkDotNet.Attributes;

namespace System.Collections
{
    public class AddGivenSize
    {
        private int[] _uniqueValues;

        public int Size = 512;

        [GlobalSetup]
        public void Setup() => _uniqueValues = Enumerable.Range(0, Size).ToArray();

        [Benchmark]
        public List<int> List()
        {
            var collection = new List<int>(Size);
            var uniqueValues = _uniqueValues;
            for (int i = 0; i < uniqueValues.Length; i++)
                collection.Add(uniqueValues[i]);
            return collection;
        }
    }
}

EgorBot · 2024-07-29T16:45:30Z

Benchmark results on Arm64

BenchmarkDotNet v0.13.12, Ubuntu 22.04.4 LTS (Jammy Jellyfish)
Unknown processor
  Job-LBCZEO : .NET 9.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD
  Job-DFVRWT : .NET 9.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD

Method	Toolchain	Mean	Error	Ratio
List	Main	1.759 μs	0.0269 μs	1.00
List	PR	1.332 μs	0.0130 μs	0.76

BDN_Artifacts.zip

Flame graphs: Main vs PR 🔥
Hot asm: Main vs PR
Hot functions: Main vs PR

For clean perf results, make sure you have just one [Benchmark] in your app.

jakobbotsch · 2024-07-29T17:08:59Z

@EgorBot -arm64 --disasm --envvars "DOTNET_JitDisasm:List" -commit 15af764 vs previous

// Licensed to the .NET Foundation under one or more agreements.
// The .NET Foundation licenses this file to you under the MIT license.
// See the LICENSE file in the project root for more information.

using BenchmarkDotNet.Attributes;

namespace System.Collections
{
    public class AddGivenSize
    {
        private int[] _uniqueValues;

        public int Size = 512;

        [GlobalSetup]
        public void Setup() => _uniqueValues = Enumerable.Range(0, Size).ToArray();

        [Benchmark]
        public List<int> List()
        {
            var collection = new List<int>(Size);
            var uniqueValues = _uniqueValues;
            for (int i = 0; i < uniqueValues.Length; i++)
                collection.Add(uniqueValues[i]);
            return collection;
        }
    }
}

jakobbotsch · 2024-07-29T17:10:24Z

@EgorBot -arm64 -commit 15af764 vs previous --disasm --envvars "DOTNET_JitDisasm:List"

// Licensed to the .NET Foundation under one or more agreements.
// The .NET Foundation licenses this file to you under the MIT license.
// See the LICENSE file in the project root for more information.

using BenchmarkDotNet.Attributes;

namespace System.Collections
{
    public class AddGivenSize
    {
        private int[] _uniqueValues;

        public int Size = 512;

        [GlobalSetup]
        public void Setup() => _uniqueValues = Enumerable.Range(0, Size).ToArray();

        [Benchmark]
        public List<int> List()
        {
            var collection = new List<int>(Size);
            var uniqueValues = _uniqueValues;
            for (int i = 0; i < uniqueValues.Length; i++)
                collection.Add(uniqueValues[i]);
            return collection;
        }
    }
}

EgorBot · 2024-07-29T17:41:09Z

Benchmark results on Arm64

BenchmarkDotNet v0.13.12, Ubuntu 22.04.4 LTS (Jammy Jellyfish)
Unknown processor
  Job-IHZRBM : .NET 9.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD
  Job-YYSFCL : .NET 9.0.0 (42.42.42.42424), Arm64 RyuJIT AdvSIMD
EnvironmentVariables=DOTNET_JitDisasm=List

Method	Toolchain	Mean	Error	Ratio	Code Size
List	Main	1.696 μs	0.0296 μs	1.00	304 B
List	PR	1.258 μs	0.0088 μs	0.74	308 B

BDN_Artifacts.zip

…dr -> ldp Very targeted fix for dotnet#93401 and dotnet#101437: before reordering two indirections, check if there is a potential store in the same loop that looks like it could end up being a candidate for store-to-load forwarding into one of those indirections. Some hardware does not handle store-to-load forwarding with the same fidelity when `stp`/`ldp` is involved compared to multiple `str`/`ldr`. If we detect the situation then avoid doing the reordering.

performanceautofiler bot added arch-arm64 os-linux Linux OS (any supported distro) runtime-coreclr specific to the CoreCLR runtime untriaged New issue has not been triaged by the area owner labels Oct 12, 2023

performanceautofiler bot mentioned this issue Oct 12, 2023

[SENTINEL] Autofile run complete at 10/12/2023 7:50:22 AM. 6 issues filed. dotnet/perf-autofiling-issues#23006

Closed

EgorBo transferred this issue from dotnet/perf-autofiling-issues Oct 12, 2023

dotnet-issue-labeler bot added the needs-area-label An area label is needed to ensure this gets routed to the appropriate area owners label Oct 12, 2023

EgorBo added area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI and removed needs-area-label An area label is needed to ensure this gets routed to the appropriate area owners labels Oct 12, 2023

EgorBo added tenet-performance Performance related issue tenet-performance-benchmarks Issue from performance benchmark labels Oct 12, 2023

jakobbotsch self-assigned this Oct 12, 2023

jakobbotsch added this to the 9.0.0 milestone Oct 12, 2023

ghost removed the untriaged New issue has not been triaged by the area owner label Oct 12, 2023

jakobbotsch added the Priority:2 Work that is important, but not critical for the release label May 3, 2024

This comment was marked as resolved.

Sign in to view

jakobbotsch mentioned this issue Jul 30, 2024

JIT: Check for potential store-to-load forwarding before reordering ldr -> ldp #105695

Merged

jakobbotsch closed this as completed in #105695 Aug 1, 2024

jakobbotsch closed this as completed in 69c9bf8 Aug 1, 2024

dotnet-policy-service bot added the in-pr There is an active PR which will close this issue when it is merged label Aug 1, 2024

github-actions bot locked and limited conversation to collaborators Sep 1, 2024

[Perf] Linux/arm64: 8 Regressions on 10/7/2023 2:03:18 PM #93401

[Perf] Linux/arm64: 8 Regressions on 10/7/2023 2:03:18 PM #93401

Comments

performanceautofiler bot commented Oct 12, 2023 • edited Loading

Run Information

Regressions in System.Collections.IterateFor<String>

Repro

Payloads

System.Collections.IterateFor<String>.ReadOnlySpan(Size: 512)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Text.Perf_Ascii

Repro

Payloads

System.Text.Perf_Ascii.ToUpper_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

JIT Disasms

System.Text.Perf_Ascii.ToLower_Bytes_Chars(Size: 128)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Buffers.Text.Tests.Base64Tests

Repro

Payloads

System.Buffers.Text.Tests.Base64Tests.ConvertToBase64CharArray(NumberOfBytes: 1000)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Collections.CreateAddAndClear<Int32>

Repro

Payloads

System.Collections.CreateAddAndClear<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Collections.Perf_Frozen<NotKnownComparable>

Repro

Payloads

System.Collections.Perf_Frozen<NotKnownComparable>.Contains_True(Count: 4)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Memory.Span<Char>

Repro

Payloads

System.Memory.Span<Char>.IndexOfAnyThreeValues(Size: 512)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

Run Information

Regressions in System.Collections.AddGivenSize<Int32>

Repro

Payloads

System.Collections.AddGivenSize<Int32>.List(Size: 512)

ETL Files

Histogram

Description of detection logic

JIT Disasms

Docs

EgorBo commented Oct 12, 2023

performanceautofiler bot commented Oct 12, 2023 •

edited

Loading