[llvm] [NVPTX] Generalize and extend upsizing when lowering 8/16-bit-element vector loads/stores (PR #119622)

[Drew Kersnar via llvm-commits]

dakersnar wrote:

Actually, there's one more thing I want to test, hold off for a little I'll let you know when it's ready.

https://github.com/llvm/llvm-project/pull/119622