[Perf] Linux/x64: 20 Regressions on 6/20/2024 7:04:47 AM #104022
Comments
performanceautofiler
bot
added
arch-x64
os-linux
Run Information
Regressions in System.Numerics.Tests.Perf_Quaternion
ReproGeneral Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md Repro StepsPrerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
Linux # Set $RunDir to the runtime directory
RunDir=`pwd`
# Set the OS, arch, and OSId
RunOS='linux'
RunOSId='linux'
RunArch='x64'
# Create aot directory
mkdir -p $RunDir/artifacts/bin/aot/sgen
mkdir -p $RunDir/artifacts/bin/aot/pack
cp -r $RunDir/artifacts/obj/mono/$RunOS.$RunArch.Release/mono/* $RunDir/artifacts/bin/aot/sgen
cp -r $RunDir/artifacts/bin/microsoft.netcore.app.runtime.$RunOS-$RunArch/Release/* $RunDir/artifacts/bin/aot/pack
# Create Core Root
$RunDir/src/tests/build.sh release $RunArch generatelayoutonly /p:LibrariesConfiguration=Release
# Clone performance
git clone --branch main --depth 1 --quiet https://github.com/dotnet/performance.git $RunDir/performance
# One line run:
python3 $RunDir/performance/scripts/benchmarks_ci.py --csproj $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --incremental no --architecture $RunArch -f net8.0 --filter 'System.Numerics.Tests.Perf_Quaternion*' --bdn-artifacts $RunDir/artifacts/BenchmarkDotNet.Artifacts --bdn-arguments="--anyCategories Libraries Runtime --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir/artifacts/bin/aot/sgen/mini/mono-sgen --customruntimepack $RunDir/artifacts/bin/aot/pack --aotcompilermode llvm --logBuildOutput --generateBinLog"
# Individual Commands:
# Restore
dotnet restore $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --packages $RunDir/performance/artifacts/packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Build
dotnet build $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore /p:NuGetPackageRoot=$RunDir/performance/artifacts/packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Run
dotnet run --project $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore --no-build -- --filter System.Numerics.Tests.Perf_Quaternion* --anyCategories Libraries Runtime " --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir/artifacts/bin/aot/sgen/mini/mono-sgen --customruntimepack $RunDir/artifacts/bin/aot/pack --aotcompilermode llvm --logBuildOutput --generateBinLog " --artifacts $RunDir/artifacts/BenchmarkDotNet.Artifacts --packages $RunDir/performance/artifacts/packages --buildTimeout 1200Windows # Set $RunDir to the runtime directory
$RunDir="FullPathHere"
# Set the OS, arch, and OSId
RunOS='windows'
RunOSId='win'
RunArch='x64'
# Create aot directory
mkdir $RunDir\artifacts\bin\aot\sgen
mkdir $RunDir\artifacts\bin\aot\pack
xcopy $RunDir\artifacts\obj\mono\$RunOS.$RunArch.Release\mono $RunDir\artifacts\bin\aot\sgen\ /e /y
xcopy $RunDir\artifacts\bin\microsoft.netcore.app.runtime.$RunOSId-$RunArch\Release $RunDir\artifacts\bin\aot\pack\ /e /y
# Create Core Root
$RunDir\src\tests\build.cmd release $RunArch generatelayoutonly /p:LibrariesConfiguration=Release
# Clone performance
git clone --branch main --depth 1 --quiet https://github.com/dotnet/performance.git $RunDir\performance
# One line run:
python3 $RunDir\performance\scripts\benchmarks_ci.py --csproj $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --incremental no --architecture $RunArch -f net8.0 --filter 'System.Numerics.Tests.Perf_Quaternion*' --bdn-artifacts $RunDir\artifacts\BenchmarkDotNet.Artifacts --bdn-arguments="--anyCategories Libraries Runtime --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir\artifacts\bin\aot\sgen\mini\mono-sgen.exe --customruntimepack $RunDir\artifacts\bin\aot\pack --aotcompilermode llvm --logBuildOutput --generateBinLog"
# Individual Commands:
# Restore
dotnet restore $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --packages $RunDir\performance\artifacts\packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Build
dotnet build $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore /p:NuGetPackageRoot=$RunDir\performance\artifacts\packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Run
dotnet run --project $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore --no-build -- --filter System.Numerics.Tests.Perf_Quaternion* --anyCategories Libraries Runtime " --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir\artifacts\bin\aot\sgen\mini\mono-sgen.exe --customruntimepack $RunDir\artifacts\bin\aot\pack -aotcompilermode llvm --logBuildOutput --generateBinLog " --artifacts $RunDir\artifacts\BenchmarkDotNet.Artifacts --packages $RunDir\performance\artifacts\packages --buildTimeout 1200System.Numerics.Tests.Perf_Quaternion.MultiplyByQuaternionOperatorBenchmarkETL FilesHistogramJIT DisasmsSystem.Numerics.Tests.Perf_Quaternion.LerpBenchmarkETL FilesHistogramJIT DisasmsSystem.Numerics.Tests.Perf_Quaternion.ConcatenateBenchmarkETL FilesHistogramJIT DisasmsSystem.Numerics.Tests.Perf_Quaternion.DivisionOperatorBenchmarkETL FilesHistogramJIT DisasmsDocsProfiling workflow for dotnet/runtime repository Run Information
Regressions in System.Numerics.Tests.Perf_Vector3
ReproGeneral Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md Repro StepsPrerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
Linux # Set $RunDir to the runtime directory
RunDir=`pwd`
# Set the OS, arch, and OSId
RunOS='linux'
RunOSId='linux'
RunArch='x64'
# Create aot directory
mkdir -p $RunDir/artifacts/bin/aot/sgen
mkdir -p $RunDir/artifacts/bin/aot/pack
cp -r $RunDir/artifacts/obj/mono/$RunOS.$RunArch.Release/mono/* $RunDir/artifacts/bin/aot/sgen
cp -r $RunDir/artifacts/bin/microsoft.netcore.app.runtime.$RunOS-$RunArch/Release/* $RunDir/artifacts/bin/aot/pack
# Create Core Root
$RunDir/src/tests/build.sh release $RunArch generatelayoutonly /p:LibrariesConfiguration=Release
# Clone performance
git clone --branch main --depth 1 --quiet https://github.com/dotnet/performance.git $RunDir/performance
# One line run:
python3 $RunDir/performance/scripts/benchmarks_ci.py --csproj $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --incremental no --architecture $RunArch -f net8.0 --filter 'System.Numerics.Tests.Perf_Vector3*' --bdn-artifacts $RunDir/artifacts/BenchmarkDotNet.Artifacts --bdn-arguments="--anyCategories Libraries Runtime --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir/artifacts/bin/aot/sgen/mini/mono-sgen --customruntimepack $RunDir/artifacts/bin/aot/pack --aotcompilermode llvm --logBuildOutput --generateBinLog"
# Individual Commands:
# Restore
dotnet restore $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --packages $RunDir/performance/artifacts/packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Build
dotnet build $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore /p:NuGetPackageRoot=$RunDir/performance/artifacts/packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Run
dotnet run --project $RunDir/performance/src/benchmarks/micro/MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore --no-build -- --filter System.Numerics.Tests.Perf_Vector3* --anyCategories Libraries Runtime " --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir/artifacts/bin/aot/sgen/mini/mono-sgen --customruntimepack $RunDir/artifacts/bin/aot/pack --aotcompilermode llvm --logBuildOutput --generateBinLog " --artifacts $RunDir/artifacts/BenchmarkDotNet.Artifacts --packages $RunDir/performance/artifacts/packages --buildTimeout 1200Windows # Set $RunDir to the runtime directory
$RunDir="FullPathHere"
# Set the OS, arch, and OSId
RunOS='windows'
RunOSId='win'
RunArch='x64'
# Create aot directory
mkdir $RunDir\artifacts\bin\aot\sgen
mkdir $RunDir\artifacts\bin\aot\pack
xcopy $RunDir\artifacts\obj\mono\$RunOS.$RunArch.Release\mono $RunDir\artifacts\bin\aot\sgen\ /e /y
xcopy $RunDir\artifacts\bin\microsoft.netcore.app.runtime.$RunOSId-$RunArch\Release $RunDir\artifacts\bin\aot\pack\ /e /y
# Create Core Root
$RunDir\src\tests\build.cmd release $RunArch generatelayoutonly /p:LibrariesConfiguration=Release
# Clone performance
git clone --branch main --depth 1 --quiet https://github.com/dotnet/performance.git $RunDir\performance
# One line run:
python3 $RunDir\performance\scripts\benchmarks_ci.py --csproj $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --incremental no --architecture $RunArch -f net8.0 --filter 'System.Numerics.Tests.Perf_Vector3*' --bdn-artifacts $RunDir\artifacts\BenchmarkDotNet.Artifacts --bdn-arguments="--anyCategories Libraries Runtime --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir\artifacts\bin\aot\sgen\mini\mono-sgen.exe --customruntimepack $RunDir\artifacts\bin\aot\pack --aotcompilermode llvm --logBuildOutput --generateBinLog"
# Individual Commands:
# Restore
dotnet restore $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --packages $RunDir\performance\artifacts\packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Build
dotnet build $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore /p:NuGetPackageRoot=$RunDir\performance\artifacts\packages /p:UseSharedCompilation=false /p:BuildInParallel=false /m:1
# Run
dotnet run --project $RunDir\performance\src\benchmarks\micro\MicroBenchmarks.csproj --configuration Release --framework net8.0 --no-restore --no-build -- --filter System.Numerics.Tests.Perf_Vector3* --anyCategories Libraries Runtime " --category-exclusion-filter NoAOT NoWASM --runtimes monoaotllvm --aotcompilerpath $RunDir\artifacts\bin\aot\sgen\mini\mono-sgen.exe --customruntimepack $RunDir\artifacts\bin\aot\pack -aotcompilermode llvm --logBuildOutput --generateBinLog " --artifacts $RunDir\artifacts\BenchmarkDotNet.Artifacts --packages $RunDir\performance\artifacts\packages --buildTimeout 1200System.Numerics.Tests.Perf_Vector3.CrossBenchmarkETL FilesHistogramJIT DisasmsSystem.Numerics.Tests.Perf_Vector3.TransformByQuaternionBenchmarkETL FilesHistogramJIT DisasmsDocsProfiling workflow for dotnet/runtime repository |
matouskozak
added
area-System.Numerics
tenet-performance
|
Tagging subscribers to this area: @dotnet/area-system-numerics |
|
Could it be related to #103527 ? cc: @tannergooding @fanyang-mono related arm64 regressions: dotnet/perf-autofiling-issues#37021 |
|
Related mono-jit regressions: dotnet/perf-autofiling-issues#36935 |
jeffhandley
added
blocking-release
needs-further-triage
|
@tannergooding / @michaelgsharp - I've marked this as blocking-release for 9.0.0 to ensure we triage this further and determine if we need to make a fix before the release. |
|
Almost all of these have fully recovered thanks to #105009 There's a couple where we've recovered most but not all the cost, those happen to be the one's using the new Accelerating the new API for Mono shouldn't be difficult, but I'd leave it open to @fanyang-mono on whether I should submit a fix for .NET 9 or wait for .NET 10 to open (given the feature complete deadline is next week). I don't believe these are considered core scenarios for the primary Mono targets |
I took a look at the arm64 regression issue (dotnet/perf-autofiling-issues#37021) and some regressions got resolved (or even better then before). However, a common theme was a slight improvement yet still significantly regressed from the previous measurements e.g. Do you think this is the best we can get from these microbenchmarks currently or are there more optimizations possible for these APIs? |
Like I mentioned above, it looks like the remaining bits here are that However, there's another aspect here as well which is that these benchmarks aren't currently "well written" and are heavily influenced by constant folding the runtime can do. Ideally they would be rewritten to not be constant foldable so we're only ever testing the real execution cost of these APIs. |
|
I took a deeper look here and it wasn't actually However, I had since added basic support for constants last year in #81902. So I was able to code up #105299, which now mirrors most of the RyuJIT |
|
Unlike |
Thank you @tannergooding for taking a second look and finding this missing piece. I'll keep an eye on the measurements after you merge that PR to see what the gains are. |
|
@matouskozak, is this one measuring |
This is MonoAOT-llvm measurement. However, in some scenarios, the real configuration used might be MonoAOT-mini because of the limitations of LLVM backend (e.g. generics are handled by mini). FYI: Mini is the backend that is used for MonoJIT. |
|
Looks like many of these were fixed and then a different regression exists now, likely due to the update to LLVM 19. It's believed that these are hitting the various failures tracked by [mono] LLVM 19 build failures · Issue #105319 · dotnet/runtime (github.com) and so are falling back to the JIT implementation which isn't accelerated for x64 today |
|
Agreed, we can close this as completed since the original regressions on MonoAOT-llvm scenarios were mitigated. The new regressions will be tracked separately since they are un-related to the original change #103527. Thank you for fixing these. |
Run Information
Regressions in System.Numerics.Tests.Perf_Vector4
Test Report
Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md
Repro Steps
Prerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
runtime/artifactsor build instructions: Libraries README args:-subset libs+libs.tests -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/coreclr/$RunOS.$RunArch.Release, build instructions: CoreCLR README args:-subset clr+libs -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/mono/$RunOS.$RunArch.Release, build instructions: MONO README args:-arch $RunArch -os $RunOS -s mono+libs+host+packs -c Release /p:CrossBuild=false /p:MonoLLVMUseCxx11Abi=falseLinux
Windows
System.Numerics.Tests.Perf_Vector4.TransformByMatrix4x4Benchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Vector4.TransformVector2ByQuaternionBenchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Vector4.TransformVector2ByMatrix4x4Benchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Vector4.LerpBenchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Vector4.TransformByQuaternionBenchmark
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
Run Information
Regressions in System.Numerics.Tests.Perf_Plane
Test Report
Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md
Repro Steps
Prerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
runtime/artifactsor build instructions: Libraries README args:-subset libs+libs.tests -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/coreclr/$RunOS.$RunArch.Release, build instructions: CoreCLR README args:-subset clr+libs -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/mono/$RunOS.$RunArch.Release, build instructions: MONO README args:-arch $RunArch -os $RunOS -s mono+libs+host+packs -c Release /p:CrossBuild=false /p:MonoLLVMUseCxx11Abi=falseLinux
Windows
System.Numerics.Tests.Perf_Plane.NormalizeBenchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Plane.TransformByMatrix4x4Benchmark
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
Run Information
Regressions in PerfLabTests.LowLevelPerf
Test Report
Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md
Repro Steps
Prerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
runtime/artifactsor build instructions: Libraries README args:-subset libs+libs.tests -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/coreclr/$RunOS.$RunArch.Release, build instructions: CoreCLR README args:-subset clr+libs -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/mono/$RunOS.$RunArch.Release, build instructions: MONO README args:-arch $RunArch -os $RunOS -s mono+libs+host+packs -c Release /p:CrossBuild=false /p:MonoLLVMUseCxx11Abi=falseLinux
Windows
PerfLabTests.LowLevelPerf.SealedClassInterfaceMethod
ETL Files
Histogram
JIT Disasms
PerfLabTests.LowLevelPerf.ClassVirtualMethod
ETL Files
Histogram
JIT Disasms
PerfLabTests.LowLevelPerf.EmptyStaticFunction5Arg
ETL Files
Histogram
JIT Disasms
PerfLabTests.LowLevelPerf.GenericGenericMethod
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
Run Information
Regressions in System.Numerics.Tests.Perf_Vector2
Test Report
Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md
Repro Steps
Prerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
runtime/artifactsor build instructions: Libraries README args:-subset libs+libs.tests -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/coreclr/$RunOS.$RunArch.Release, build instructions: CoreCLR README args:-subset clr+libs -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/mono/$RunOS.$RunArch.Release, build instructions: MONO README args:-arch $RunArch -os $RunOS -s mono+libs+host+packs -c Release /p:CrossBuild=false /p:MonoLLVMUseCxx11Abi=falseLinux
Windows
System.Numerics.Tests.Perf_Vector2.TransformNormalByMatrix4x4Benchmark
ETL Files
Histogram
JIT Disasms
System.Numerics.Tests.Perf_Vector2.TransformNormalByMatrix3x2Benchmark
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
Run Information
Regressions in System.Collections.Sort<IntStruct>
Test Report
Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md
Repro Steps
Prerequisites (Files either built locally (with build.(sh/cmd) or downloaded from payload above (if same system setup) (in this order))
runtime/artifactsor build instructions: Libraries README args:-subset libs+libs.tests -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/coreclr/$RunOS.$RunArch.Release, build instructions: CoreCLR README args:-subset clr+libs -rc release -configuration Release -arch $RunArch -framework net8.0runtime/artifacts/bin/mono/$RunOS.$RunArch.Release, build instructions: MONO README args:-arch $RunArch -os $RunOS -s mono+libs+host+packs -c Release /p:CrossBuild=false /p:MonoLLVMUseCxx11Abi=falseLinux
Windows
System.Collections.Sort<IntStruct>.Array(Size: 512)
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
The text was updated successfully, but these errors were encountered: