[clang] [Clang] VectorExprEvaluator::VisitCallExpr / InterpretBuiltin - Allow GFNI intrinsics to be used in constexpr (PR #169619)

[NagaChaitanya Vellanki via cfe-commits]

chaitanyav wrote:

@RKSimon  I extracted the common functions(`GFNIAffine`, `GFNIMul`, `GFNIMultiplicativeInverse`) used in both (new, current) interpreter and made them available to both via `ExprConstShared.h` (based on @efriedma-quic feedback on my other PR). I am also using a multiplicative inverse lookup table to speed up the inverse affine. Please let me know if anything else can be simplified/optimized. Named the variables so that the implementation is closer to the intel intrinsic guide.

https://github.com/llvm/llvm-project/pull/169619