Massively simplify getBaseJitTypeAndSizeOfSIMDType and the general SIMD handle caching/lookup

[tannergooding]

This resolves #83091

-- Wanting to get TP numbers to see if the getBaseJitTypeAndSizeOfSIMDType changes in particular are a win. I'm expecting them to be so because we have, on average, less overall checks being done for what used to be both the fast and slow path.

[ghost]

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

Issue Details

---

This resolves #83091

-- Wanting to get TP numbers to see if the getBaseJitTypeAndSizeOfSIMDType changes in particular are a win. I'm expecting them to be so because we have, on average, less overall checks being done for what used to be both the fast and slow path.

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

[tannergooding]

CC. @jakobbotsch, @dotnet/jit-contrib

This is effectively no change to TP, with there being a very small win on coreclr_tests (-0.01%) for x64.

Massively simplifies the handle lookup and caching logic. -- It could likely have the total LoC reduced even more, if desired, but I didn't feel like creating a CLASS_HANDLE* and then double indirecting to avoid repeated code between the various TYP_SIMD apis...

Also fixes a couple of bugs that existed in the code that caused invalid handles to be returned and so resolves the general issue in 83091. -- There is a separate issue in that a couple gtNewSimdHWIntrinsic calls specify the wrong TYP_SIMD, but I will be fixing that in a follow up PR with a different assert.

[kunalspathak]

LGTM. Nice cleanup.