[Perf] Windows/x64 Regressions in System.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>
performanceautofiler opened this issue · comments
performanceautofiler commented
Run Information
Name | Value |
---|---|
Architecture | x64 |
OS | Windows 10.0.22621 |
Queue | TigerWindows |
Baseline | 1d622621b6bc254fbd777921c70a3f5d57fbd6b8 |
Compare | a71a85a65f6527ca2754ccf336f1fc77f02c409b |
Diff | Diff |
Configs | CompilationMode:tiered, RunKind:micro |
Regressions in System.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>
Benchmark | Baseline | Test | Test/Base | Test Quality | Edge Detector | Baseline IR | Compare IR | IR Ratio |
---|---|---|---|---|---|---|---|---|
1.49 μs | 4.46 μs | 2.99 | 0.70 | False | ||||
5.38 μs | 6.11 μs | 1.13 | 0.32 | False | ||||
76.12 ns | 240.82 ns | 3.16 | 0.54 | False |
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.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>*'
System.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>.Round(BufferLength: 3079)
ETL Files
Histogram
JIT Disasms
System.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>.Sin(BufferLength: 3079)
ETL Files
Histogram
JIT Disasms
System.Numerics.Tensors.Tests.Perf_FloatingPointTensorPrimitives<Double>.Round(BufferLength: 128)
ETL Files
Histogram
JIT Disasms
Docs
Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository
Parker Bibus commented
Seems to likely be caused by: #102098 with clear regression across multiple platforms in ADX. @tannergooding
Commit range: 1d62262...29cb8a6
dotnet-policy-service commented
Tagging subscribers to this area: @dotnet/area-system-numerics
See info in area-owners.md if you want to be subscribed.
Drew Scoggins commented
Related: dotnet/perf-autofiling-issues#34775