search

dotnet / perf-autofiling-issues

A landing place for auto-filed performance issues before they receive triage
MIT License
9 stars 4 forks source link

[Perf] Linux/x64: 20 Improvements on 8/18/2024 5:03:53 PM #40106

Open performanceautofiler[bot] opened 3 weeks ago

performanceautofiler[bot] commented 3 weeks ago

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in System.IO.Tests.Perf_FileStream

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
106.84 ms 80.14 ms 0.75 0.01 True
107.98 ms 81.94 ms 0.76 0.01 True
1.34 ms 1.08 ms 0.81 0.02 False
1.59 ms 1.19 ms 0.75 0.20 False
1.05 ms 760.69 μs 0.73 0.01 True
1.11 ms 900.64 μs 0.81 0.04 True

graph graph graph graph graph graph 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*'
### System.IO.Tests.Perf_FileStream.Write_NoBuffering_PreallocationSize(fileSize: 104857600, userBufferSize: 16384, options: None) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_FileStream.WriteAsync_NoBuffering_PreallocationSize(fileSize: 104857600, userBufferSize: 16384, options: Asynchronous) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_FileStream.Append(fileSize: 1048576, userBufferSize: 4096, options: DeleteOnClose) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_FileStream.AppendAsync(fileSize: 1048576, userBufferSize: 4096, options: DeleteOnClose, Asynchronous) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_FileStream.Write_NoBuffering_PreallocationSize(fileSize: 1048576, userBufferSize: 16384, options: None) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_FileStream.WriteAsync_NoBuffering_PreallocationSize(fileSize: 1048576, userBufferSize: 16384, options: Asynchronous) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in System.IO.Tests.Perf_File

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
27.54 μs 22.30 μs 0.81 0.13 False
14.78 μs 12.58 μs 0.85 0.14 False
142.98 μs 115.33 μs 0.81 0.02 False
26.51 μs 21.31 μs 0.80 0.04 False
20.39 μs 16.30 μs 0.80 0.39 False
112.09 ms 89.09 ms 0.79 0.02 True
1.07 ms 788.91 μs 0.74 0.00 True
49.80 μs 46.06 μs 0.92 0.05 False

graph graph graph graph graph graph graph graph 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*'
### System.IO.Tests.Perf_File.AppendAllTextAsync(size: 10000) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.AppendAllTextAsync(size: 100) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.WriteAllText(size: 100000) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.AppendAllText(size: 10000) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.AppendAllLinesAsync #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.CopyTo(size: 104857600) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.CopyTo(size: 1048576) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Tests.Perf_File.WriteAllText(size: 10000) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in System.Collections.IndexerSet<Int32>

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
6.52 μs 5.89 μs 0.90 0.02 False

graph 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.IndexerSet&lt;Int32&gt;*'
### System.Collections.IndexerSet<Int32>.SortedDictionary(Size: 512) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in BenchmarksGame.FannkuchRedux_9

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
90.99 ms 78.17 ms 0.86 0.00 True

graph 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 'BenchmarksGame.FannkuchRedux_9*'
### BenchmarksGame.FannkuchRedux_9.RunBench(n: 11, expectedSum: 556355) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in BenchmarksGame.FannkuchRedux_5

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
15.40 ms 13.40 ms 0.87 0.02 False

graph 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 'BenchmarksGame.FannkuchRedux_5*'
### BenchmarksGame.FannkuchRedux_5.RunBench(n: 10, expectedSum: 38) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in System.IO.Tests.Perf_Directory

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
4.30 ms 3.91 ms 0.91 0.01 False

graph 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*'
### System.IO.Tests.Perf_Directory.EnumerateFiles #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)

Run Information

Name Value
Architecture x64
OS ubuntu 22.04
Queue ViperUbuntu
Baseline bb1b43a6a155eaeb45953a3c87f6e04b12979778
Compare d3da7a2c823245534e557f97013136d06a98e731
Diff Diff
Configs CompilationMode:tiered, RunKind:micro

Improvements in System.IO.Pipes.Tests.Perf_NamedPipeStream

Benchmark Baseline Test Test/Base Test Quality Edge Detector Baseline IR Compare IR IR Ratio
183.72 μs 152.28 μs 0.83 0.14 False
184.25 μs 146.00 μs 0.79 0.09 False

graph graph 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.Pipes.Tests.Perf_NamedPipeStream*'
### System.IO.Pipes.Tests.Perf_NamedPipeStream.ReadWrite(size: 1000000, Options: Asynchronous) #### ETL Files #### Histogram #### JIT Disasms ### System.IO.Pipes.Tests.Perf_NamedPipeStream.ReadWrite(size: 1000000, Options: None) #### ETL Files #### Histogram #### JIT Disasms ### Docs [Profiling workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/profiling-workflow-dotnet-runtime.md) [Benchmarking workflow for dotnet/runtime repository](https://github.com/dotnet/performance/blob/master/docs/benchmarking-workflow-dotnet-runtime.md)
AndyAyersMS commented 3 weeks ago

Only diff in range is https://github.com/dotnet/runtime/pull/106578.

@jkotas FYI

AndyAyersMS commented 3 weeks ago

Also https://github.com/dotnet/perf-autofiling-issues/issues/40040

lewing commented 3 weeks ago

Only diff in range is dotnet/runtime#106578.

@jkotas FYI

That change unblocked the official build against the new helix image with the new windows sdk, so this would be the first perf run against the new images

LoopedBard3 commented 1 day ago

This is only being seen on Ubuntu x64 machines, maybe there was an Ubuntu update of some sort causing this? We do run apt update so there is a chance it was caused by that. Though a quick look over the logs doesn't show any obvious updates in the runs before the improvement.