[Perf] Windows/x64: 3 Regressions on 9/1/2026 5:28:08 AM +00:00

### Run Information

Name | Value
-- | --
Architecture | x64
OS | Windows 10.0.22631
Queue | ViperWindows
Baseline | [ef3efe5e020e412e4f6453781d59443e975010a4](https://github.com/dotnet/runtime/commit/ef3efe5e020e412e4f6453781d59443e975010a4)
Compare | [818ad35699cf9b1a3e3f2067c133d8dbe599da2e](https://github.com/dotnet/runtime/commit/818ad35699cf9b1a3e3f2067c133d8dbe599da2e)
Diff | [Diff](https://github.com/dotnet/runtime/compare/ef3efe5e020e412e4f6453781d59443e975010a4...818ad35699cf9b1a3e3f2067c133d8dbe599da2e)
Configs | CompilationMode:tiered, RunKind:micro

### Regressions in System.Numerics.Tests.Perf_BigInteger

Benchmark | Baseline | Test | Test/Base | Test Quality | Edge Detector | Baseline IR | Compare IR | IR Ratio
-- | -- | -- | -- | -- | -- | -- | -- | --
|

• [Multiply - Duration of single invocation]()
• 📝 - [Benchmark Source]()
• [ADX - Test Multi Config Graph]()

| 6.50 ns | 12.74 ns | 1.96 | 0.07 | False | | |
|

• [Multiply - Duration of single invocation]()
• 📝 - [Benchmark Source]()
• [ADX - Test Multi Config Graph]()

| 6.50 ns | 12.73 ns | 1.96 | 0.10 | False | | |
|

• [ModPow - Duration of single invocation]()
• 📝 - [Benchmark Source]()
• [ADX - Test Multi Config Graph]()

| 81.28 ns | 91.56 ns | 1.13 | 0.02 | False | | |

![graph]()
![graph]()
![graph]()
[Test Report]()

### Repro
General Docs link: https://github.com/dotnet/performance/blob/main/docs/benchmarking-workflow-dotnet-runtime.md

```cmd
git clone https://github.com/dotnet/performance.git
py .\performance\scripts\benchmarks_ci.py -f net8.0 --filter 'System.Numerics.Tests.Perf_BigInteger*'
```

### System.Numerics.Tests.Perf_BigInteger.Multiply(arguments: 16,16 bits)

#### ETL Files

#### Histogram

#### JIT Disasms

### System.Numerics.Tests.Perf_BigInteger.Multiply(arguments: 16,8 bits)

#### ETL Files

#### Histogram

#### JIT Disasms

### System.Numerics.Tests.Perf_BigInteger.ModPow(arguments: 16,16,16 bits)

#### 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)

Start with the benchmark definitions in performance/src/benchmarks/micro/libraries/System.Runtime.Numerics/Perf.BigInteger.cs at the cited Multiply and ModPow lines. Run the provided benchmarks_ci.py command, then compare the baseline commit ef3efe5e with 818ad356 to identify the source of the Windows x64 regression. Done means the affected benchmarks no longer reproduce the reported slowdown and the performance report confirms the improvement.

Written by the indexing model from the issue text.