[Perf] Windows/x64: 16 Regressions on 8/28/2024 6:34:28 PM

[performanceautofiler]

Run Information

Name	Value
Architecture	x64
OS	Windows 10.0.22621
Queue	TigerWindows
Baseline	7ff684af49cdf008bebbe10efcc10a2649795d65
Compare	063ae05c63307becef004eefbc8704ccb03c7f04
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in System.IO.Compression.Brotli

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
Decompress_WithoutState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	514.85 μs	574.60 μs	1.12	0.00	False
Decompress - Duration of single invocation ADX - Test Multi Config Graph	170.38 μs	185.91 μs	1.09	0.00	False
Decompress_WithState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	158.93 μs	174.27 μs	1.10	0.00	False
Decompress_WithoutState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	347.69 μs	509.89 μs	1.47	0.01	True
Decompress_WithoutState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	158.30 μs	173.88 μs	1.10	0.00	False
Decompress - Duration of single invocation ADX - Test Multi Config Graph	429.39 μs	594.83 μs	1.39	0.01	False
Decompress_WithoutState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	600.56 μs	677.33 μs	1.13	0.01	False
Decompress_WithState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	606.75 μs	678.74 μs	1.12	0.00	False
Decompress_WithState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	516.26 μs	574.34 μs	1.11	0.00	False
Decompress_WithoutState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	370.87 μs	544.47 μs	1.47	0.01	False
Decompress - Duration of single invocation ADX - Test Multi Config Graph	399.41 μs	554.34 μs	1.39	0.01	True
Decompress - Duration of single invocation ADX - Test Multi Config Graph	574.80 μs	632.68 μs	1.10	0.00	False
Decompress - Duration of single invocation ADX - Test Multi Config Graph	662.45 μs	733.76 μs	1.11	0.00	False
Decompress_WithState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	371.45 μs	544.75 μs	1.47	0.01	False
Decompress_WithState - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	343.18 μs	509.99 μs	1.49	0.01	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
py .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.IO.Compression.Brotli*'

System.IO.Compression.Brotli.Decompress_WithoutState(level: Optimal, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress(level: Fastest, file: "sum")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithState(level: Fastest, file: "sum")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithoutState(level: Fastest, file: "TestDocument.pdf")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithoutState(level: Fastest, file: "sum")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress(level: Optimal, file: "TestDocument.pdf")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithoutState(level: Fastest, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithState(level: Fastest, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithState(level: Optimal, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithoutState(level: Optimal, file: "TestDocument.pdf")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress(level: Fastest, file: "TestDocument.pdf")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress(level: Optimal, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress(level: Fastest, file: "alice29.txt")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithState(level: Optimal, file: "TestDocument.pdf")

ETL Files

Histogram

JIT Disasms

System.IO.Compression.Brotli.Decompress_WithState(level: Fastest, file: "TestDocument.pdf")

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	Windows 10.0.22621
Queue	TigerWindows
Baseline	7ff684af49cdf008bebbe10efcc10a2649795d65
Compare	063ae05c63307becef004eefbc8704ccb03c7f04
Diff	Diff
Configs	CompilationMode:tiered, RunKind:micro

Regressions in PerfLabTests.CastingPerf2.CastingPerf

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio
ObjrefValueTypeObj - Duration of single invocation 📝 - Benchmark Source ADX - Test Multi Config Graph	1.06 ms	1.27 ms	1.19	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
py .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'PerfLabTests.CastingPerf2.CastingPerf*'

PerfLabTests.CastingPerf2.CastingPerf.ObjrefValueTypeObj

ETL Files

Histogram

JIT Disasms

Docs

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

[dotnet-policy-service]

Tagging subscribers to this area: @dotnet/area-system-io-compression
See info in area-owners.md if you want to be subscribed.

[LoopedBard3]

Caused by: #106994 (Update of Brotli library) @jkoritzinsky
Also on x86: dotnet/perf-autofiling-issues#40619

Other regressions:
Linux x64: dotnet/perf-autofiling-issues#40578, dotnet/perf-autofiling-issues#40570
Linux arm64: dotnet/perf-autofiling-issues#40700

We also saw some improvements:

• [Perf] Windows/x64: 2 Improvements on 8/28/2024 6:34:28 PM perf-autofiling-issues#40625
• [Perf] Linux/x64: 7 Improvements on 8/28/2024 6:34:28 PM perf-autofiling-issues#40581

[ericstj]

@jkoritzinsky @carlossanlop - it seems these regressions are due to the brotli update. What's your take on this, do we resolve as by-design or should someone take a closer look and share the findings upstream?

[carlossanlop]

Related discussion here: dotnet/perf-autofiling-issues#40587

[carlossanlop]

I opened an issue in google/brotli with our results to try to understand the regressions: google/brotli#1202

[carlossanlop]

Since we are taking this brotli update anyway, I think we can close this issue. The discussion can continue in the issue I opened in the google/brotli repo.