Remove the rest of the SimdAsHWIntrinsic support

[tannergooding]

This removes quite a bit of complexity and ensures that Vector<T> handling is identical to the handling for the fixed sized vector types on exposed APIs.

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

[tannergooding]

This one is for .NET 10 only, it does not need backport

[tannergooding]

/azp run runtime-coreclr jitstress-isas-x86, jitstress-isas-arm64, Fuzzlyn

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[tannergooding]

CC. @dotnet/jit-contrib This is ready for review.

This removes the last of the handling that was specific to System.Numerics.* namespace and fully unifies it with the handling in the System.Runtime.Intrinsics.* namespace by mapping across the APIs during import.

This results in some fairly significant size and throughput wins for both x64 and Arm64, mostly stemming from places that the System.Numerics.* handling wasn't going down the same code paths and lighting up the same scenarios: https://dev.azure.com/dnceng-public/public/_build/results?buildId=856735&view=ms.vss-build-web.run-extensions-tab

[EgorBo]

Nice clean up! I've noticed some missing ternary ops (avx512) among the regressions but I guess it's not a big deal

[tannergooding]

I've noticed some missing ternary ops (avx512) among the regressions but I guess it's not a big deal

Yeah, we've got a tracking issue for those: #107619 and #90075

We get a lot of the cases, but there's a bit of slightly more advanced folding we still want to handle

[kunalspathak]

LGTM