[LoopedBard3]

Summary

Move desktop main measurements to SDK 12 and net12.0, rather than building net11 benchmarks with an SDK-only installation that contains only the .NET 12 runtime.

• Update the main channel, root SDK pin, runtime pipeline default, supported microbenchmark frameworks, SDK scenario inventory, Blazor package selection, tooling test targets, and ResultsComparer framework recognition.
• Keep MAUI/mobile main-channel targets on net11.0 with SDK 12 until their dependencies support net12. Separate framework selection from SDK acquisition using ci_setup.py --framework, retaining MAUI/macios dependency lookups on their existing net11.0 branches instead of nonexistent net12 branches.
• Preserve explicitly versioned older channels, SDK selection, package versions, and their existing BenchmarkDotNet runner. Add global.net11.json alongside the existing older SDK configurations.
• Use BenchmarkDotNet 0.16.0-nightly.20261014.649 for net12. The previous 0.16.0-nightly.20260801.595 restores/builds net12 but crashes before execution with GetRuntimeVersion not implemented for NotRecognized. Scope the new runner to net12 because its runtime/toolchain API changes would otherwise affect older Mono/WASM measurements.
• Retain AOT exclusion semantics with a tested compatibility attribute and an explicit async-LINQ dependency for the new runner. Make TLS/URI single-argument sources strongly typed for its analyzer without changing benchmark names, argument values, or measured methods.
• Select corewasm12.0 for net12 CoreCLR WASM. Reject unsupported net12 Mono WASM explicitly instead of silently selecting a different measurement target.

This change is independent of the MAUI ACES pool migration and contains no pool changes. Based on freshly fetched dotnet/performance main at bee760ce03eda6021a1466486e067c7c7e76a4ef.

Validation

Executed on Windows x64 with this checkout's private tools\dotnet\x64 first on PATH; PYTHONPATH was set to this checkout's scripts directory for pytest. $privateSdk below denotes the absolute path returned by (Resolve-Path tools\dotnet\x64).Path.

Command	Actual result
python -m pytest scripts\tests\test_channel_map.py scripts\tests\test_dotnet.py scripts\tests\test_run_performance_job.py scripts\tests\test_blazor_workload.py scripts\tests\test_wasm_coreclr_r2r.py -q --basetemp=artifacts\net12-pytest-final	167 passed
dotnet test src\tests\harness\BenchmarkDotNet.Extensions.Tests\BenchmarkDotNet.Extensions.Tests.csproj --configuration Release --framework net12.0 -p:PERFLAB_TARGET_FRAMEWORKS=net12.0 --nologo --verbosity minimal	56 passed, including AOT exclusion tests and full real-data partition enumeration
dotnet test src\tools\ResultsComparer.Tests\ResultsComparer.Tests.csproj --configuration Release --framework net12.0 --nologo --verbosity minimal	21 passed
dotnet test src\tools\Reporting\Reporting.Tests\Reporting.Tests.csproj --configuration Release --framework net12.0 --nologo --verbosity minimal	13 passed
python scripts\ci_setup.py --channel main --architecture x64 --not-in-lab	Passed; acquired SDK 12.0.100-alpha.1.26527.106 and runtime 12.0.0-alpha.1.26527.106, selected net12.0
python scripts\benchmarks_ci.py --frameworks net12.0 --dotnet-path $privateSdk --filter '*Perf_Guid.NewGuid*' --bdn-arguments '--job Dry'	Restore, build, and execution passed; 1 benchmark executed. Both host and benchmark process reported runtime 12.0.0-alpha.1.26527.106, SDK 12.0.100-alpha.1.26527.106, and BDN 0.16.0-nightly.20261014.649. The repository-configured job was used, not a literal Dry job.
python scripts\benchmarks_ci.py --frameworks net11.0 --dotnet-path $privateSdk --build-only	Passed, 0 warnings / 0 errors, retaining BDN 0.16.0-nightly.20260801.595
python scripts\ci_setup.py --channel main --framework net11.0 --architecture x64 --not-in-lab --install-dir $privateSdk --output-file tools\maui-net11-setup	Passed; generated DOTNET_VERSION=12.0.100-alpha.1.26527.106, PERFLAB_TARGET_FRAMEWORKS=net11.0, and PERFLAB_BRANCH=12.0
git diff --check	Passed

The private installation contained only .NET 12 runtimes during the successful net12 execution, so the benchmark was not a net11 benchmark implicitly rolled forward. The smoke run is execution evidence, not a performance comparison.

Remaining validation limits

• Full MAUI workload acquisition, application builds, and execution on Android/iOS devices were not run. SDK/target separation, emitted setup values, and dependency branch selection are covered by focused tests and a real setup invocation.
• Actual WASM workload build/execution was not run; argument selection and the existing WASM/CoreCLR/R2R regression suite passed. Net12 Mono WASM remains unsupported and is rejected explicitly; older targets retain their existing runner/arguments.
• The complete Azure Pipelines/Helix matrix, CertHelper tests, and Startup tests were not run locally.
• The older net11 control was restored/built, not executed; no .NET 11 runtime was added to the isolated installation.