[Perf] Changes at 5/13/2021 9:46:09 PM

[performanceautofiler]

Run Information

Architecture	arm64
OS	Windows 10.0.19041
Baseline	d77854a16eaa29f1a141f1da267885dcf2675027
Compare	509ed7ae74de659dc5a8d38696ae103b70ba82a1
Diff	Diff

Regressions in Microsoft.Extensions.DependencyInjection.TimeToFirstService

Benchmark	Baseline	Test	Test/Base	Test Quality	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
Singleton - Duration of single invocation	3.71 μs	4.24 μs	1.14	0.31
BuildProvider - Duration of single invocation	1.30 μs	1.43 μs	1.10	0.34
Transient - Duration of single invocation	3.64 μs	4.18 μs	1.15	0.21
Transient - Duration of single invocation	3.64 μs	4.60 μs	1.26	0.28
Scoped - Duration of single invocation	4.15 μs	4.68 μs	1.13	0.29
Transient - Duration of single invocation	3.63 μs	4.13 μs	1.14	0.27
Scoped - Duration of single invocation	4.20 μs	4.73 μs	1.13	0.26
Singleton - Duration of single invocation	3.72 μs	4.31 μs	1.16	0.26
Singleton - Duration of single invocation	3.70 μs	4.23 μs	1.15	0.25
Singleton - Duration of single invocation	3.66 μs	4.23 μs	1.15	0.28
Scoped - Duration of single invocation	4.15 μs	4.65 μs	1.12	0.24
Transient - Duration of single invocation	3.65 μs	4.17 μs	1.14	0.22

Historical Data in Reporting System

Repro

git clone https://github.com/dotnet/performance.git
py .\performance\scripts\benchmarks_ci.py -f netcoreapp5.0 --filter 'Microsoft.Extensions.DependencyInjection.TimeToFirstService*'

Payloads

Baseline
Compare

Histogram

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Singleton(Mode: "Runtime")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.BuildProvider(Mode: "Expressions")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Transient(Mode: "Runtime")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Transient(Mode: "ILEmit")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Scoped(Mode: "Dynamic")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Transient(Mode: "Expressions")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Scoped(Mode: "ILEmit")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Singleton(Mode: "Expressions")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Singleton(Mode: "Dynamic")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Singleton(Mode: "ILEmit")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Scoped(Mode: "Runtime")

Microsoft.Extensions.DependencyInjection.TimeToFirstService.Transient(Mode: "Dynamic")

Docs

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

### Run Information

Architecture	arm64
OS	Windows 10.0.19041
Baseline	d77854a16eaa29f1a141f1da267885dcf2675027
Compare	509ed7ae74de659dc5a8d38696ae103b70ba82a1
Diff	Diff

Regressions in System.Collections.CopyTo<Int32>

Benchmark	Baseline	Test	Test/Base	Test Quality	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
Span - Duration of single invocation	167.92 ns	189.12 ns	1.13	0.19

Historical Data in Reporting System

Repro

git clone https://github.com/dotnet/performance.git
py .\performance\scripts\benchmarks_ci.py -f netcoreapp5.0 --filter 'System.Collections.CopyTo<Int32>*'

Payloads

Baseline
Compare

Histogram

System.Collections.CopyTo<Int32>.Span(Size: 2048)

Docs

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

[DrewScoggins]

dotnet/runtime#52691