[llvm] [NVPTX] Vectorize and lower 256-bit global loads/stores for sm_100+/ptx88+ (PR #139292)

[Drew Kersnar via llvm-commits]

dakersnar wrote:

Do we want the above formatting change? I figured it was laid out the way it is for a reason, but maybe that reason is just "it was written before we started checking formatting" 😀.

https://github.com/llvm/llvm-project/pull/139292