[Perf] Changes at 2/8/2022 9:06:59 PM

[performanceautofiler]

Run Information

Architecture	x64
OS	ubuntu 18.04
Baseline	75589aeb707aad7247a7215ff42bb92d94eb6e22
Compare	205f70e04b5f57b54a97967c1e2c3f198519540d
Diff	Diff

Improvements in System.Tests.Perf_Double

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
Parse - Duration of single invocation	298.53 ns	105.31 ns	0.35	0.01	True
TryParse - Duration of single invocation	306.85 ns	108.50 ns	0.35	0.00	True

Test Report

Repro

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.Tests.Perf_Double*'

Payloads

Baseline
Compare

Histogram

System.Tests.Perf_Double.Parse(value: "1.7976931348623157e+308")

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 105.31192177898758 < 284.35895249408327.
IsChangePoint: Marked as a change because one of 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 662.1976453764302 (T) = (0 -106.23664456578346) / Math.Sqrt((16.968903039471314 / (299)) + (0.6689632359725259 / (26))) is greater than 1.967335607330539 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (299) + (26) - 2, .975) and 0.6415975174335248 = (296.417155944446 - 106.23664456578346) / 296.417155944446 is greater than 0.05.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

```#### System.Tests.Perf_Double.TryParse(value: "-1.7976931348623157e+308")

```log

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 108.50413388900739 < 290.468315734174.
IsChangePoint: Marked as a change because one of 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 545.5482291337471 (T) = (0 -107.84369387179426) / Math.Sqrt((20.69160196559321 / (189)) + (0.27129803399050184 / (26))) is greater than 1.971163885465162 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (189) + (26) - 2, .975) and 0.6365954592531325 = (296.7593460724358 - 107.84369387179426) / 296.7593460724358 is greater than 0.05.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

Docs

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

### Run Information

Architecture	x64
OS	ubuntu 18.04
Baseline	75589aeb707aad7247a7215ff42bb92d94eb6e22
Compare	205f70e04b5f57b54a97967c1e2c3f198519540d
Diff	Diff

Improvements in System.IO.Tests.Perf_StreamWriter

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
WritePartialCharArray - Duration of single invocation	114.31 ms	89.81 ms	0.79	0.05	False

Test Report

Repro

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.IO.Tests.Perf_StreamWriter*'

Payloads

Baseline
Compare

Histogram

System.IO.Tests.Perf_StreamWriter.WritePartialCharArray(writeLength: 2)

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 89.80775440000001 < 106.30712238708334.
IsChangePoint: Marked as a change because one of 2/3/2022 10:14:29 PM, 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 40.59909458700451 (T) = (0 -89478174.12435545) / Math.Sqrt((14291641554714.633 / (299)) + (957361075625.5273 / (26))) is greater than 1.967335607330539 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (299) + (26) - 2, .975) and 0.11659854166382454 = (101288234.56198646 - 89478174.12435545) / 101288234.56198646 is greater than 0.05.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

Docs

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

Run Information

Architecture	x64
OS	ubuntu 18.04
Baseline	75589aeb707aad7247a7215ff42bb92d94eb6e22
Compare	205f70e04b5f57b54a97967c1e2c3f198519540d
Diff	Diff

Improvements in System.Threading.Tests.Perf_CancellationToken

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
CreateLinkedTokenSource2 - Duration of single invocation	65.59 ns	59.75 ns	0.91	0.03	False

Test Report

Repro

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.Threading.Tests.Perf_CancellationToken*'

Payloads

Baseline
Compare

Histogram

System.Threading.Tests.Perf_CancellationToken.CreateLinkedTokenSource2

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 59.74737834063577 < 62.425235025047606.
IsChangePoint: Marked as a change because one of 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 32.70998784700469 (T) = (0 -60.45420314356094) / Math.Sqrt((1.3737078843850912 / (299)) + (0.8555414797249012 / (26))) is greater than 1.967335607330539 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (299) + (26) - 2, .975) and 0.09484040784765856 = (66.7884466647582 - 60.45420314356094) / 66.7884466647582 is greater than 0.05.
IsChangeEdgeDetector: Marked not as a regression because Edge Detector said so.

Docs

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

Run Information

Architecture	x64
OS	ubuntu 18.04
Baseline	75589aeb707aad7247a7215ff42bb92d94eb6e22
Compare	205f70e04b5f57b54a97967c1e2c3f198519540d
Diff	Diff

Improvements in System.Buffers.Text.Tests.Utf8ParserTests

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
TryParseSingle - Duration of single invocation	110.05 ns	58.59 ns	0.53	0.02	True
TryParseDouble - Duration of single invocation	262.91 ns	73.90 ns	0.28	0.02	True

Test Report

Repro

git clone https://github.com/dotnet/performance.git
python3 .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.Buffers.Text.Tests.Utf8ParserTests*'

Payloads

Baseline
Compare

Histogram

System.Buffers.Text.Tests.Utf8ParserTests.TryParseSingle(value: 3.4028235E+38)

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 58.59341631134521 < 105.27264513827608.
IsChangePoint: Marked as a change because one of 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 468.75334950533124 (T) = (0 -58.736844070929706) / Math.Sqrt((3.5620520116527943 / (299)) + (0.044088413215427245 / (26))) is greater than 1.967335607330539 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (299) + (26) - 2, .975) and 0.48213131865157605 = (113.42034416522543 - 58.736844070929706) / 113.42034416522543 is greater than 0.05.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

```#### System.Buffers.Text.Tests.Utf8ParserTests.TryParseDouble(value: -1.7976931348623157e+308)

```log

Description of detection logic

IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsRegressionBase: Marked as not a regression because the compare was not 5% greater than the baseline, or the value was too small.
IsImprovementBase: Marked as improvement because the compare was 5% less than the baseline, and the value was not too small.
IsImprovementCheck: Marked as improvement because the three check build points were 0.05 less than the baseline.
IsImprovementWindowed:Marked as improvement because 73.8998664043469 < 251.3822589367363.
IsChangePoint: Marked as a change because one of 2/8/2022 5:20:12 PM, 2/15/2022 5:17:31 AM falls between 2/6/2022 2:48:36 PM and 2/15/2022 5:17:31 AM.
IsImprovementStdDev: Marked as improvement because 731.5443395322558 (T) = (0 -72.37170628583785) / Math.Sqrt((10.474015373478244 / (299)) + (0.9299225633190419 / (26))) is greater than 1.967335607330539 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (299) + (26) - 2, .975) and 0.7289632449971537 = (267.01805179551314 - 72.37170628583785) / 267.01805179551314 is greater than 0.05.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

Docs

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

[stephentoub]

Seems unlikely the CancellationToken improvements were related to double/float parsing?

Similarly for StreamWriter?

[EgorBo]

Seems unlikely the CancellationToken improvements were related to double/float parsing?

Similarly for StreamWriter?

Yeah I forgot to mentions those, we looked at them and came to conclusion only your "rollup !!" could improve them dotnet/runtime#64720

[danmoseley]

Yeah I forgot to mentions those, we looked at them and came to conclusion only your "rollup !!" could improve them dotnet/runtime#64720

How did it do that?

[EgorBo]

Yeah I forgot to mentions those, we looked at them and came to conclusion only your "rollup !!" could improve them dotnet/runtime#64720

How did it do that?

Example - https://sharplab.io/#v2:EYLgtghglgdgNAFxBAzmAPgAQEwEYCwAUJgMwAEOZAwmQN5FmNkNNMuOYAsZAcgKYB3ABQB5YACs+AYwRkA9gEIFASnZ01rTLgCcQ5QG41AXzVquZEQBsAJqInTZc1YVb0XrJlABmZIXLIAvAFkMACulpbOHtFkCAAWAE5yAiGCZACCCQDmoWB8MAg84ZYAogAeUnwADghQcjBCMBB5cl5+ygYabO7RWrqdPWQmg2bcAJIwlrB8ACp8KAh2kjLyUUxuMRY27fpkAPR7ZAnVlhCVZAJQ8byCO8ZERkA==

!! operator makes code way smaller and inliner-friendly

[danmoseley]

Interesting. It essentially follows the throw-helper pattern we already use in various places, including reuse of the throw-helper.