Ensure that the CpuId tests set preferredVectorByteLength to a non-zero value

[tannergooding]

This resolves#88581 and resolves#88582 and resolves#88580

[ghost]

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

Issue Details

---

This resolves #88581

Author:	tannergooding
Assignees:	-
Labels:	area-CodeGen-coreclr
Milestone:	-

[tannergooding]

/azp run runtime-coreclr jitstress-isas-x86, runtime-coreclr outerloop, runtime-coreclr jitstress, runtime-coreclr libraries-jitstress

[azure-pipelines]

Azure Pipelines successfully started running 4 pipeline(s).

[tannergooding]

/azp run runtime-coreclr jitstress-isas-x86, runtime-coreclr outerloop, runtime-coreclr jitstress, runtime-coreclr libraries-jitstress

[azure-pipelines]

Azure Pipelines successfully started running 4 pipeline(s).

[tannergooding]

/azp run runtime-coreclr jitstress-isas-x86, runtime-coreclr outerloop, runtime-coreclr jitstress, runtime-coreclr libraries-jitstress

[azure-pipelines]

No commit pushedDate could be found for PR 88623 in repo dotnet/runtime

[tannergooding]

/azp run runtime-coreclr jitstress-isas-x86, runtime-coreclr outerloop, runtime-coreclr jitstress, runtime-coreclr libraries-jitstress

[azure-pipelines]

Azure Pipelines successfully started running 4 pipeline(s).

[tannergooding]

CC. @dotnet/jit-contrib. This resolves #88581 and resolves #88582 and resolves #88580

The two main issues were that PreferredVectorByteWidth was being tied to the jit-stress mode, which made it non-deterministic and differ per-method. This broke cases like Vector512.IsHardwareAccelerated.

Additionally, the check around IsHardwareAccelerated itself wasn't properly conditioned and could differ in some scenarios, leading to other issues and inconsistencies.

This fixes both of those issues and ensures the tests are taking the preferred width into account so they pass as expected.

[AndyAyersMS]

Seems like the stress mode might be salvageable somehow, but let's sort that out later.

[tannergooding]

Seems like the stress mode might be salvageable somehow, but let's sort that out later.

It's worth noting that we are explicitly testing PreferredVectorBitWidth=512 in stress already, so it wasn't buying us much.

[tannergooding]

@dotnet-policy-service rerun

[tannergooding]

There are still some stress failures, but unrelated to this PR in particular.