[Perf] Linux/x64: 79 Improvements on 6/25/2026 3:50:04 PM +00:00

[performanceautofiler]

Run Information

Name	Value
Architecture	x64
OS	ubuntu 22.04
Queue	ViperUbuntu
Baseline	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.IO.Tests.Perf_FileStream

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
GetLength - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2151.04 ns	834.36 ns	0.39	0.17	False
GetLength - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2138.92 ns	831.10 ns	0.39	0.15	False
SeekBackward - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2312.37 μs	927.08 μs	0.40	0.21	False
SeekBackward - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2313.38 μs	936.58 μs	0.40	0.14	False
LockUnlock - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.07 μs	7.66 μs	0.69	0.22	False
LockUnlock - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.03 μs	7.72 μs	0.70	0.20	False
OpenClose - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.36 μs	5.25 μs	0.71	0.21	False
OpenClose - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	7.29 μs	5.25 μs	0.72	0.19	False
Read - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.97 μs	7.02 μs	0.78	0.21	False
Read - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.94 μs	7.02 μs	0.79	0.16	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.IO.Tests.Perf_FileStream*'

Details

System.IO.Tests.Perf_FileStream.GetLength(options: Read)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.GetLength(options: Write)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.SeekBackward(fileSize: 1024, options: Asynchronous)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.SeekBackward(fileSize: 1024, options: None)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.LockUnlock(fileSize: 1024, options: Asynchronous)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.LockUnlock(fileSize: 1024, options: None)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.OpenClose(fileSize: 1024, options: None)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.OpenClose(fileSize: 1024, options: Asynchronous)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.Read(fileSize: 1024, userBufferSize: 1024, options: None)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_FileStream.Read(fileSize: 1024, userBufferSize: 1024, options: Asynchronous)

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.IO.Tests.Perf_File

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ReadAllBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.81 μs	7.67 μs	0.56	0.15	False
ReadAllBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.27 μs	7.52 μs	0.57	0.20	False
ReadAllBytes - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.57 μs	8.90 μs	0.61	0.18	False
CopyToOverwrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	21.31 μs	15.23 μs	0.71	0.14	False
CopyToOverwrite - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	21.85 μs	15.66 μs	0.72	0.21	False
CopyTo - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	30.21 μs	23.40 μs	0.77	0.17	False
AppendAllText - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.76 μs	6.85 μs	0.78	0.15	False
CopyTo - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	30.85 μs	24.20 μs	0.78	0.21	False
Exists - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.87 μs	2.40 μs	0.84	0.17	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.IO.Tests.Perf_File*'

Details

System.IO.Tests.Perf_File.ReadAllBytes(size: 4096)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.ReadAllBytes(size: 512)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.ReadAllBytes(size: 16384)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.CopyToOverwrite(size: 512)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.CopyToOverwrite(size: 4096)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.CopyTo(size: 512)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.AppendAllText(size: 100)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.CopyTo(size: 4096)

ETL Files

Histogram

JIT Disasms

System.IO.Tests.Perf_File.Exists

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
MultiplyBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	23.24 ns	13.06 ns	0.56	0.41	False
MultiplyOperatorBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.81 ns	13.73 ns	0.60	0.34	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.00 ns	10.65 ns	0.76	0.55	False
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.25 ns	10.12 ns	0.76	0.48	False
DotBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	56.48 ns	45.52 ns	0.81	0.22	False
GreaterThanOrEqualBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.20 ns	10.70 ns	0.81	0.47	False
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.86 ns	12.40 ns	0.83	0.51	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.MultiplyBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.MultiplyOperatorBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.DotBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.GreaterThanOrEqualBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.AbsBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<SByte>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
MultiplyBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.65 ns	13.10 ns	0.58	0.37	False
MultiplyOperatorBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.90 ns	14.49 ns	0.63	0.40	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.88 ns	9.82 ns	0.76	0.60	False
MinBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.91 ns	10.75 ns	0.90	0.47	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.Numerics.Tests.Perf_VectorOf<SByte>*'

Details

System.Numerics.Tests.Perf_VectorOf<SByte>.MultiplyBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<SByte>.MultiplyOperatorBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<SByte>.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<SByte>.MinBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	15.39 ns	11.36 ns	0.74	0.53	False
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.26 ns	9.92 ns	0.75	0.51	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.46 ns	10.82 ns	0.87	0.55	False
EqualsStaticBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.90 ns	10.42 ns	0.88	0.47	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>.AbsBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt64>.EqualsStaticBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<SByte>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
MultiplyBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	23.36 ns	13.19 ns	0.56	0.36	False
MultiplyOperatorBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.90 ns	13.42 ns	0.59	0.34	False
DotBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	57.72 ns	46.79 ns	0.81	0.21	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<SByte>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<SByte>.MultiplyBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<SByte>.MultiplyOperatorBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<SByte>.DotBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<Byte>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
MultiplyBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.98 ns	12.85 ns	0.56	0.36	False
MultiplyOperatorBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	22.68 ns	13.09 ns	0.58	0.36	False
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	15.07 ns	12.44 ns	0.83	0.51	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.Numerics.Tests.Perf_VectorOf<Byte>*'

Details

System.Numerics.Tests.Perf_VectorOf<Byte>.MultiplyBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<Byte>.MultiplyOperatorBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<Byte>.AbsBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_Vector3

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
CreateFromScalar - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.33 ns	10.65 ns	0.74	0.44	False
UnitYBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.43 ns	10.01 ns	0.75	0.52	False
UnitXBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.33 ns	10.90 ns	0.82	0.52	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.Numerics.Tests.Perf_Vector3*'

Details

System.Numerics.Tests.Perf_Vector3.CreateFromScalar

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_Vector3.UnitYBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_Vector3.UnitXBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Double>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.14 ns	10.23 ns	0.78	0.54	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.87 ns	10.99 ns	0.79	0.59	False
GreaterThanBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.76 ns	10.51 ns	0.89	0.48	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Double>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Double>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Double>.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Double>.GreaterThanBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Memory.Span<Byte>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1295.95 ns	101.61 ns	0.08	0.01	False
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	273.68 ns	76.13 ns	0.28	0.12	False
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	78.32 ns	68.68 ns	0.88	0.18	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<Byte>*'

Details

System.Memory.Span<Byte>.CopyToOverlap(Size: 3000)

ETL Files

Histogram

JIT Disasms

System.Memory.Span<Byte>.CopyToOverlap(Size: 512)

ETL Files

Histogram

JIT Disasms

System.Memory.Span<Byte>.CopyToOverlap(Size: 33)

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Memory.Span<Char>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2523.60 ns	127.32 ns	0.05	0.00	False
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	493.15 ns	83.05 ns	0.17	0.01	False
CopyToOverlap - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	93.53 ns	69.10 ns	0.74	0.17	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<Char>*'

Details

System.Memory.Span<Char>.CopyToOverlap(Size: 3000)

ETL Files

Histogram

JIT Disasms

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

ETL Files

Histogram

JIT Disasms

System.Memory.Span<Char>.CopyToOverlap(Size: 33)

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.IO.MemoryMappedFiles.Tests.Perf_MemoryMappedFile

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
CreateFromFile_Read - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.62 μs	9.24 μs	0.73	0.15	False
CreateFromFile_Read - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.58 μs	11.10 μs	0.76	0.20	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.IO.MemoryMappedFiles.Tests.Perf_MemoryMappedFile*'

Details

System.IO.MemoryMappedFiles.Tests.Perf_MemoryMappedFile.CreateFromFile_Read(capacity: 10000)

ETL Files

Histogram

JIT Disasms

System.IO.MemoryMappedFiles.Tests.Perf_MemoryMappedFile.CreateFromFile_Read(capacity: 100000)

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Single>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.85 ns	10.08 ns	0.73	0.52	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.04 ns	11.63 ns	0.89	0.56	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Single>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Single>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Single>.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int16>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.69 ns	10.30 ns	0.75	0.52	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.80 ns	10.09 ns	0.79	0.53	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int16>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int16>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int16>.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Float

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.89 ns	11.17 ns	0.80	0.51	False
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.39 ns	11.00 ns	0.82	0.53	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.Runtime.Intrinsics.Tests.Perf_Vector128Float*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Float.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Float.AllBitsSetBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<UInt16>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	15.91 ns	12.19 ns	0.77	0.51	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.34 ns	10.53 ns	0.79	0.56	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.Numerics.Tests.Perf_VectorOf<UInt16>*'

Details

System.Numerics.Tests.Perf_VectorOf<UInt16>.AbsBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<UInt16>.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<UInt32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	15.75 ns	11.52 ns	0.73	0.52	False
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.36 ns	11.46 ns	0.86	0.62	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.Numerics.Tests.Perf_VectorOf<UInt32>*'

Details

System.Numerics.Tests.Perf_VectorOf<UInt32>.AbsBenchmark

ETL Files

Histogram

JIT Disasms

System.Numerics.Tests.Perf_VectorOf<UInt32>.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.98 ns	11.18 ns	0.75	0.56	False
MultiplyOperatorBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	11.89 ns	10.52 ns	0.89	0.44	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt32>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt32>.AbsBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt32>.MultiplyOperatorBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt16>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.49 ns	9.67 ns	0.72	0.52	False
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.85 ns	11.26 ns	0.81	0.55	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt16>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt16>.AllBitsSetBenchmark

ETL Files

Histogram

JIT Disasms

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<UInt16>.AbsBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int64>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.97 ns	11.71 ns	0.90	0.57	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int64>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int64>.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	13.69 ns	10.51 ns	0.77	0.55	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.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int32>*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Int32>.AllBitsSetBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in Benchstone.BenchF.Simpsn

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Test - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	416.59 ms	371.51 ms	0.89	0.05	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 'Benchstone.BenchF.Simpsn*'

Details

Benchstone.BenchF.Simpsn.Test

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in MicroBenchmarks.Serializers.Xml_FromStream<CollectionsOfPrimitives>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
DataContractSerializer_ - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	19.42 ms	13.25 ms	0.68	0.19	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 'MicroBenchmarks.Serializers.Xml_FromStream<CollectionsOfPrimitives>*'

Details

MicroBenchmarks.Serializers.Xml_FromStream<CollectionsOfPrimitives>.DataContractSerializer_

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<UInt64>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AbsBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.90 ns	11.22 ns	0.75	0.47	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.Numerics.Tests.Perf_VectorOf<UInt64>*'

Details

System.Numerics.Tests.Perf_VectorOf<UInt64>.AbsBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Runtime.Intrinsics.Tests.Perf_Vector128Int

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
AllBitsSetBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	14.14 ns	8.85 ns	0.63	0.53	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.Runtime.Intrinsics.Tests.Perf_Vector128Int*'

Details

System.Runtime.Intrinsics.Tests.Perf_Vector128Int.AllBitsSetBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.IO.Tests.Perf_Directory

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Exists - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	2.74 μs	2.36 μs	0.86	0.18	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.IO.Tests.Perf_Directory*'

Details

System.IO.Tests.Perf_Directory.Exists

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_Vector2

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	8.75 ns	6.53 ns	0.75	0.60	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.Numerics.Tests.Perf_Vector2*'

Details

System.Numerics.Tests.Perf_Vector2.ZeroBenchmark

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Tests.Perf_DateTimeOffset

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
op_Subtraction - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	17.24 ns	15.01 ns	0.87	0.23	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.Tests.Perf_DateTimeOffset*'

Details

System.Tests.Perf_DateTimeOffset.op_Subtraction

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Memory.Constructors<String>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
SpanImplicitCastFromArray - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	20.14 ns	18.07 ns	0.90	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.Memory.Constructors<String>*'

Details

System.Memory.Constructors<String>.SpanImplicitCastFromArray

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	818cfd5edcda78ac3eb201af392fd301ab9ad46b
Compare	4e28c96c3bd32ce2811ef7128641e8926874020e
Diff	Diff
Configs	CompilationMode:wasm, RunKind:micro

Improvements in System.Numerics.Tests.Perf_VectorOf<Int64>

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ZeroBenchmark - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	12.34 ns	11.20 ns	0.91	0.54	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.Numerics.Tests.Perf_VectorOf<Int64>*'

Details

System.Numerics.Tests.Perf_VectorOf<Int64>.ZeroBenchmark

ETL Files

Histogram

JIT Disasms

Docs

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

[lewing]

ps://github.com/dotnet/runtime/pull/129299

[LoopedBard3]

🔍 Automated Triage Analysis

Summary: Broad improvement across 79 tests (66 confirmed improvements, 9 borderline, 4 noise) with deltas ranging from -9% to -95%, primarily attributed to commit 4e28c96c3bd3 which enabled Vector128 fast paths on Wasm via PackedSimd. The most dramatic gains are in Span CopyToOverlap tests (-72% to -95%) and FileStream operations (-22% to -61%). Finding confidence 4/5 (excellent historical data, 295-300 runs per test), improvement confidence 5/5 (surgical wasm-specific change, extreme z-scores up to -338 Bisection skipped — autofile run configuration is not supported by the bisection harness.

Finding Confidence: 5/5 (analysis accuracy)
Improvement Confidence: 5/5 (likelihood of true improvement)

🤖 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, bimodal-undermines-confidence)

📊 Full Analysis (click to expand)

Summary: Broad improvement across 79 tests (66 confirmed improvements, 9 borderline, 4 noise) with deltas ranging from -9% to -95%, primarily attributed to commit 4e28c96c3bd3 which enabled Vector128 fast paths on Wasm via PackedSimd. The most dramatic gains are in Span CopyToOverlap tests (-72% to -95%) and FileStream operations (-22% to -61%). Finding confidence 4/5 (excellent historical data, 295-300 runs per test), improvement confidence 5/5 (surgical wasm-specific change, extreme z-scores up to -338, directly relevant to WASM+Mono stack).

Issue Overview

• Issue: #75301 - [Perf] Linux/x64: 79 Improvements on 6/25/2026 3:50:04 PM +00:00
• Type: Performance Improvement
• Date: 2026-06-25 | Queue: Ubuntu.2204.Amd64.Viper.Perf | OS/Arch: linux/x64 (wasm compilation)
• Commits: 818cfd5edcda → 4e28c96c3bd3 (17 commits)
• Labels: untriaged, perf-improvement, arch-x64, arch-wasm, os-linux, kind-micro, compilationmode-wasm, runkind-micro
• Runtime Stack: WASM + Mono
• Historical Data: Available (295-300 runs per test, excellent baselines)

Test Analysis

Span/Memory Tests (Most Dramatic Improvements)

Test Name	Baseline	Compare	Δ%	μ	σ	n	z	Thresh%	Noise?	Assessment
Span<Char>.CopyToOverlap(3000)	2523.60	127.32	-94.95%	2530.0	7.09	297	-338.1	5.0%	N	Step change, ultra-stable
Span<Byte>.CopyToOverlap(3000)	1295.95	101.61	-92.16%	1297.0	6.38	299	-187.3	5.0%	N	Step change, ultra-stable
Span<Char>.CopyToOverlap(512)	493.15	83.05	-83.16%	494.0	4.31	298	-95.0	5.0%	N	Step change, ultra-stable
Span<Byte>.CopyToOverlap(512)	273.68	76.13	-72.18%	274.0	4.69	298	-42.2	5.0%	N	Step change, ultra-stable
Span<Char>.CopyToOverlap(33)	93.53	69.10	-26.12%	93.5	3.97	298	-6.16	5.0%	N	Step change
Span<Byte>.CopyToOverlap(33)	78.32	68.68	-12.31%	78.3	4.15	299	-2.32	6.1%	Borderline	Borderline step

FileStream Tests

Test Name	Baseline	Compare	Δ%	μ	σ	n	z	Thresh%	Noise?	Assessment
GetLength(Read)	2151.04	834.36	-61.21%	2211.6	127.2	300	-10.8	11.5%	N	Step change
GetLength(Write)	2138.92	831.10	-61.14%	2139.0	90.6	295	-14.4	5.0%	N	Step change
SeekBackward(Async)	2312.37	927.08	-59.91%	2312.0	176.5	298	-7.84	7.8%	N	Step change
SeekBackward(None)	2313.38	936.58	-59.51%	2313.0	127.9	298	-10.8	5.6%	N	Step change
LockUnlock(Async)	11.07	7.66	-30.81%	11.1	0.75	298	-4.55	7.2%	N	Step change
LockUnlock(None)	11.03	7.72	-30.05%	11.0	0.74	299	-4.50	7.1%	N	Step change
OpenClose(None)	7.36	5.25	-28.58%	7.36	0.57	296	-3.67	8.3%	N	Step change
OpenClose(Async)	7.29	5.25	-28.07%	7.29	0.46	300	-4.41	6.8%	N	Step change
Read(None)	8.97	7.02	-21.72%	8.97	0.65	299	-2.98	8.1%	N	Step change
Read(Async)	8.94	7.02	-21.48%	8.94	0.65	299	-2.96	8.2%	N	Step change

File Tests

Test Name	Baseline	Compare	Δ%	μ	σ	n	z	Thresh%	Noise?	Assessment
ReadAllBytes(4096)	13.81	7.67	-44.47%	13.8	0.70	298	-8.80	5.1%	N	Step change
ReadAllBytes(512)	13.27	7.52	-43.37%	13.3	0.70	300	-8.20	5.5%	N	Step change
ReadAllBytes(16384)	14.57	8.90	-38.88%	14.6	0.63	298	-9.00	5.0%	N	Step change
CopyToOverwrite(512)	21.31	15.23	-28.53%	21.3	0.76	298	-7.98	5.0%	N	Step change
CopyToOverwrite(4096)	21.85	15.66	-28.33%	21.9	1.12	298	-5.50	6.3%	N	Step change
CopyTo(512)	30.21	23.40	-22.55%	30.2	1.16	298	-5.89	5.0%	N	Step change
AppendAllText(100)	8.76	6.85	-21.82%	8.76	0.53	298	-3.65	6.8%	N	Step change
CopyTo(4096)	30.85	24.20	-21.56%	30.9	1.24	298	-5.37	5.0%	N	Step change
Exists	2.87	2.40	-16.45%	2.87	0.14	298	-3.37	5.0%	N	Step change

Vector128/Numerics Tests (Representative Selection)

Test Name	Baseline	Compare	Δ%	μ	σ	n	z	Thresh%	Noise?	Assessment
Vector128Of<Byte>.Multiply	23.24	13.06	-43.81%	23.2	1.46	298	-6.98	6.5%	N	Step change
Vector128Of<SByte>.Multiply	23.36	13.19	-43.55%	23.4	1.27	298	-8.04	5.5%	N	Step change
VectorOf<Byte>.Multiply	22.98	12.85	-44.06%	23.0	1.45	298	-6.97	6.6%	N	Step change
VectorOf<SByte>.Multiply	22.65	13.10	-42.17%	22.7	1.19	299	-8.02	5.6%	N	Step change
Vector128Of<SByte>.MultiplyOp	22.90	13.42	-41.39%	22.9	1.39	296	-6.81	6.4%	N	Step change
Vector128Of<Byte>.MultiplyOp	22.81	13.73	-39.81%	22.8	1.29	297	-7.06	6.0%	N	Step change
Vector128Int.AllBitsSet	14.14	8.85	-37.44%	14.1	1.41	298	-3.75	9.8%	N	Step change
VectorOf<SByte>.MultiplyOp	22.90	14.49	-36.73%	22.9	1.20	298	-7.02	5.5%	N	Step change
Xml_FromStream DataContract	19.42	13.25	-31.78%	19.4	1.81	297	-3.40	10.0%	N	Step change
Vector128Of<UInt16>.AllBitsSet	13.49	9.67	-28.32%	13.5	1.25	297	-3.06	10.0%	N	Step change
VectorOf<UInt32>.Abs	15.75	11.52	-26.87%	15.7	1.02	298	-4.17	5.8%	N	Step change
Vector128Of<UInt64>.Abs	15.39	11.36	-26.15%	15.4	1.12	298	-3.60	6.9%	N	Step change
Vector3.CreateFromScalar	14.33	10.65	-25.72%	14.3	1.41	300	-2.60	8.5%	N	Step change
Vector3.UnitY	13.43	10.01	-25.47%	13.4	1.37	299	-2.49	10.7%	N	Step change
VectorOf<UInt64>.Abs	14.90	11.22	-24.74%	14.9	0.72	297	-5.12	5.1%	N	Step change
Vector2.Zero	8.75	6.53	-25.46%	8.75	0.85	298	-2.63	9.3%	N	Step change
DateTimeOffset.op_Subtraction	17.24	15.01	-12.92%	17.2	0.76	298	-2.93	5.5%	N	Step change
Simpsn.Test	416.59	371.51	-10.82%	416.6	11.57	295	-3.90	5.0%	N	Step change
Directory.Exists	2.74	2.36	-13.97%	2.74	0.13	297	-2.85	5.6%	N	Step change

Borderline Tests

Test Name	Baseline	Compare	Δ%	z	Noise?	Assessment
Vector128Of<Byte>.AllBitsSet	13.25	10.12	-23.62%	-2.40	Borderline	Near-threshold, moderate variance
Vector128Of<UInt64>.EqualsStatic	11.90	10.42	-12.38%	-1.70	Borderline	Below z threshold
Vector128Of<Double>.AllBitsSet	13.14	10.23	-22.17%	-2.00	Borderline	High CV (13.6%)
Span<Byte>.CopyToOverlap(33)	78.32	68.68	-12.31%	-2.32	Borderline	Near threshold
Vector128Of<Single>.Zero	13.04	11.63	-10.80%	-2.20	Borderline	Below adaptive threshold
VectorOf<UInt32>.Zero	13.36	11.46	-14.27%	-1.96	Borderline	Near z=2 boundary
Vector128Of<Int64>.Zero	12.97	11.71	-9.77%	-1.83	Borderline	Below threshold
Constructors<String>.SpanImplicit	20.14	18.07	-10.25%	-1.82	Borderline	Moderate variance
VectorOf<Int64>.Zero	12.34	11.20	-9.29%	-1.91	Borderline	Moderate CV (8.4%)

Noise Tests

Test Name	Baseline	Compare	Δ%	z	Assessment
Vector128Of<Byte>.GreaterThanOrEqual	13.20	10.70	-18.95%	-1.03	High variance (CV=14.7%), within noise
VectorOf<SByte>.Min	11.91	10.75	-9.78%	-1.27	High variance (CV=11.7%)
Vector128Of<Double>.GreaterThan	11.76	10.51	-10.62%	-1.31	High variance (CV=12.8%)
Vector128Of<UInt32>.MultiplyOp	11.89	10.52	-11.49%	-1.29	High variance (CV=12.6%)

Legend: μ=mean, σ=stddev, n=runs, z=z-score, Thresh%=adaptive threshold

Finding Confidence: 4/5 - Excellent historical data (295-300 runs per test), strong statistical power. Only 4 after-runs available but values are tightly clustered. Downgraded from 5 due to limited after-sample and no verification via RunWithBuildAtHash.

Candidate Stack Relevance

Commit	Touched Paths (summary)	Relevance	Notes
4e28c96c3bd3	src/libraries/System.Private.CoreLib/src/System/Runtime/Intrinsics/Vector128.cs, HexConverter.cs, Guid.cs, Utf8Utility.*.cs, TeddyHelper.cs, ProbabilisticMap.cs, Adler32.cs, XxHashShared.cs	Relevant	Directly enables Vector128 PackedSimd fast paths on Wasm — the primary beneficiary of WASM+Mono stack
2e1bb6526a1f	eng/, src/mono/browser/, src/mono/CMakeLists.txt, Emscripten version bump	Relevant	Upgrades Emscripten 5.0.6 & LLVM 23 — may contribute to general wasm codegen improvements
d08d3aca601f	src/mono/mono/metadata/native-library.c	Relevant	Mono DllImport fix, minor
42f3669b3a7f	src/coreclr/tools/Common/JitInterface/CorInfoImpl.cs	LessLikely	CoreCLR crossgen2 only
946057658995	src/coreclr/jit/fgopt.cpp	LessLikely	CoreCLR JIT only
69417eb28730	src/libraries/System.Text.Json/	Shared	BCL JSON writer, but test improvements are not in JSON area (except DataContractSerializer)

Confirmed Improvements

Category 1: Span CopyToOverlap (-12% to -95%)

Representative Test: System.Memory.Span<Char>.CopyToOverlap(Size: 3000) (-94.95%)

• Delta: 2523.60 → 127.32 (-94.95%)
• Historical Context: μ=2530.0 (σ=7.09, n=297); z-score=-338.1; adaptive threshold=5.0%
• Assessment: Step change, ultra-stable history (CV=0.3%). This is the highest-confidence improvement in the entire issue.
• Finding Confidence: 5/5 - Ultra-stable baseline with 297 runs, CV < 1%, z-score magnitude is extreme
• Improvement Confidence: 5/5 - Massive delta, surgical commit, directly relevant to WASM stack
• Likely Cause: 4e28c96c3bd3 - "[wasm] Enable Vector128 fast paths on Wasm via PackedSimd"
  • Affected Areas: Vector128 / Span / SIMD intrinsics / Wasm PackedSimd
  • Stack Relevance: Relevant
  • Rationale: This commit enables PackedSimd.IsSupported branches alongside existing Sse2/AdvSimd.Arm64 gates in Vector128.UnpackLow/UnpackHigh. The Span.CopyToOverlap operation uses Buffer.Memmove which leverages Vector128 SIMD paths — enabling PackedSimd on WASM allows the vectorized copy path to activate where previously only scalar fallback was used, explaining the dramatic 10-20x speedup on larger sizes.
  • Candidates Considered:
    • 4e28c96c3bd3 - Primary: enables Vector128 on Wasm
    • 2e1bb6526a1f - Secondary: Emscripten/LLVM upgrade may contribute
• Verification Status: Not verified via RunWithBuildAtHash. Signal is unambiguous (z=-338) and does not require verification.

Additional tests in this category:

• Span<Byte>.CopyToOverlap(3000): -92.16% (z=-187.3)
• Span<Char>.CopyToOverlap(512): -83.16% (z=-95.0)
• Span<Byte>.CopyToOverlap(512): -72.18% (z=-42.2)
• Span<Char>.CopyToOverlap(33): -26.12% (z=-6.16)

Category 2: FileStream Operations (-22% to -61%)

Representative Test: System.IO.Tests.Perf_FileStream.GetLength(options: Read) (-61.21%)

• Delta: 2151.04 → 834.36 (-61.21%)
• Historical Context: μ=2211.6 (σ=127.2, n=300); z-score=-10.8; adaptive threshold=11.5%
• Assessment: Step change. The before-series has bimodal spikes but the improvement signal is unambiguous.
• Finding Confidence: 4/5 - Excellent run count (300), clear signal despite bimodal history
• Improvement Confidence: 4/5 - Strong signal with consistent improvements across all FileStream variants
• Likely Cause: 2e1bb6526a1f - "Upgrade Emscripten to 5.0.6 & LLVM to 23" and/or 4e28c96c3bd3
  • Affected Areas: FileStream / File I/O / Wasm host / Emscripten toolchain
  • Stack Relevance: Relevant
  • Rationale: The Emscripten 5.0.6 upgrade (commit 2e1bb6526a1f) modifies src/mono/browser/ and src/libraries/System.Private.CoreLib/src/System/IO/FileStatus.Unix.cs, directly affecting Wasm file I/O. The LLVM 23 upgrade may improve codegen for I/O syscall wrappers. Additionally, the Vector128 PackedSimd enablement (4e28c96c3bd3) could accelerate buffer operations within FileStream. The consistent 22-61% improvement across all FileStream operations suggests a shared I/O infrastructure improvement.
  • Candidates Considered:
    • 2e1bb6526a1f - Emscripten/LLVM upgrade, modifies FileStatus.Unix.cs
    • 4e28c96c3bd3 - Vector128 enablement may affect buffer ops in FileStream
• Verification Status: Not verified. Bisection between these two commits would isolate the cause.

Additional tests in this category:

• GetLength(Write): -61.14% (z=-14.4)
• SeekBackward(Async): -59.91% (z=-7.84)
• SeekBackward(None): -59.51% (z=-10.8)
• LockUnlock(Async): -30.81% (z=-4.55)
• LockUnlock(None): -30.05% (z=-4.50)
• OpenClose(None): -28.58% (z=-3.67)
• OpenClose(Async): -28.07% (z=-4.41)
• Read(None): -21.72% (z=-2.98)
• Read(Async): -21.48% (z=-2.96)

Category 3: File Operations (-16% to -44%)

Representative Test: System.IO.Tests.Perf_File.ReadAllBytes(size: 4096) (-44.47%)

• Delta: 13.81 → 7.67 (-44.47%)
• Historical Context: μ=13.8 (σ=0.70, n=298); z-score=-8.80; adaptive threshold=5.1%
• Assessment: Step change with clear signal across all File operation tests.
• Finding Confidence: 4/5 - Strong historical data, consistent improvements
• Improvement Confidence: 4/5 - Multiple corroborating tests, clear step change
• Likely Cause: Same as FileStream category — Emscripten upgrade and/or Vector128 enablement
  • Stack Relevance: Relevant
  • Rationale: File.ReadAllBytes uses FileStream internally; improvements cascade from the same underlying I/O path optimizations. The vectorized buffer operations enabled by PackedSimd would benefit the internal byte-copying paths.

Category 4: Vector128/Numerics Intrinsics (-10% to -44%)

Representative Test: System.Runtime.Intrinsics.Tests.Perf_Vector128Of<Byte>.MultiplyBenchmark (-43.81%)

• Delta: 23.24 → 13.06 (-43.81%)
• Historical Context: μ=23.2 (σ=1.46, n=298); z-score=-6.98; adaptive threshold=6.5%
• Assessment: Step change. Consistent ~40-44% improvement across all Multiply variants, ~20-28% across Zero/AllBitsSet/Abs variants.
• Finding Confidence: 4/5 - Excellent historical data across all Vector tests
• Improvement Confidence: 5/5 - Directly attributable to Vector128 PackedSimd enablement commit
• Likely Cause: 4e28c96c3bd3 - "[wasm] Enable Vector128 fast paths on Wasm via PackedSimd"
  • Affected Areas: Vector128 / System.Numerics / SIMD intrinsics
  • Stack Relevance: Relevant
  • Rationale: This commit directly adds PackedSimd.IsSupported branches to Vector128.UnpackLow/UnpackHigh and related operations. The Multiply benchmarks specifically exercise Vector128.Multiply which now routes through Wasm PackedSimd hardware instructions instead of scalar fallback. The ~40% improvement in multiply tests vs ~20-25% in simpler operations (Zero, AllBitsSet) reflects the higher computational savings from vectorized arithmetic vs simple memory operations.
  • Candidates Considered:
    • 4e28c96c3bd3 - Direct: enables Vector128 SIMD paths on Wasm
    • 2e1bb6526a1f - LLVM 23 may improve SIMD codegen quality
• Verification Status: Not verified. Signal is strong and attribution is clear.

Category 5: MemoryMappedFile (-24% to -27%)

• MemoryMappedFile.CreateFromFile_Read(10000): -26.76% (z=-3.78)
• MemoryMappedFile.CreateFromFile_Read(100000): -23.87% (z=-3.45)
• Likely Cause: Emscripten upgrade improving memory-mapped file syscall wrappers

Category 6: Other Improvements

• Benchstone.BenchF.Simpsn.Test: -10.82% (z=-3.90) — Likely benefits from improved LLVM codegen
• Xml_FromStream DataContractSerializer: -31.78% (z=-3.40) — Likely benefits from improved general codegen
• DateTimeOffset.op_Subtraction: -12.92% (z=-2.93)
• Directory.Exists: -13.97% (z=-2.85) — I/O path improvement
• Vector2.Zero: -25.46% (z=-2.63) — Vector/SIMD path improvement
• Vector3.CreateFromScalar: -25.72% (z=-2.60)
• Vector3.UnitY: -25.47% (z=-2.49)
• Vector3.UnitX: -18.27% (z=-2.67)

Open Questions

• Attribution Ambiguity: Two commits likely contribute — the Emscripten/LLVM upgrade (2e1bb6526a1f) and the PackedSimd enablement (4e28c96c3bd3). Bisection between these would isolate which improvements belong to which commit.
• FileStream Mechanism: The 60% FileStream improvement is large and could stem from Emscripten's filesystem emulation improvements, LLVM codegen improvements for I/O wrappers, or indirect benefits from Vector128 enablement in buffer operations.
• Borderline Tests: 9 tests classified as borderline have deltas in the -9% to -24% range but z-scores below 2.0 or high CV. These may solidify as improvements with more after-runs.
• After-Run Count: Only 4 after-run samples per test. While the values are tightly clustered (supporting the step-change conclusion), additional runs would increase confidence.
• Bisection Status: Not performed.

Recommended Actions

1. No immediate action needed - The improvements are attributed to two intentional commits: the Wasm PackedSimd enablement (4e28c96c3bd3) and the Emscripten/LLVM upgrade (2e1bb6526a1f).
2. Monitor stability - Confirm the improvements hold across subsequent runs, especially for the borderline tests.
3. Optional bisection - To precisely attribute FileStream improvements, run bisection between commits 2e1bb6526a1f and 4e28c96c3bd3 using RunWithBuildAtHash with a representative FileStream test (e.g., Perf_FileStream.GetLength).

🤖 Proposed Next Actions

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

Bimodal benchmark history and no clean confirmed per-test subset — the bisection result for the bimodal test(s) would be unreliable, so the cause is not auto-classified as known.

#	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, bimodal-undermines-confidence, unsupported-culprit

Rationale: Every bisection was skipped so no culprit was surfaced by bisection, yet the analysis assigns culprit SHAs (4e28c96c3bd3, 2e1bb6526a1f) with 5/5 change confidence — those SHAs are unsupported guesses. The flagged bimodal history further undermines certainty that the measured shifts are true improvements rather than mode-switching artifacts; change and finding confidence should both be lowered to reflect the absence of bisection convergence and the bimodal signal.

Disputed culprit: 4e28c96c3bd3 (advisory — not applied)

Suggested confidence: Finding Confidence → 3/5; 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 2026-06-30 18:40 UTC