[llvm] [AMDGPU] Support atomic load and store for vector float types (v2f16, v2i16, v4i16, v4f16, v2f32) (PR #192904)

[Harrison Hao via llvm-commits]

================
@@ -2328,7 +2334,9 @@ defm : GlobalFLATLoadPats_D16 <GLOBAL_LOAD_SHORT_D16, load_d16_lo_global, v2f16>
 defm : GlobalFLATStorePats <GLOBAL_STORE_BYTE, atomic_store_8_global, i32>;
 defm : GlobalFLATStorePats <GLOBAL_STORE_SHORT, atomic_store_16_global, i32>;
 defm : GlobalFLATStorePats <GLOBAL_STORE_DWORD, atomic_store_32_global, i32>;
+defm : GlobalFLATStorePats <GLOBAL_STORE_DWORD, atomic_store_32_global, v2i16>;
 defm : GlobalFLATStorePats <GLOBAL_STORE_DWORDX2, atomic_store_64_global, i64>;
+defm : GlobalFLATStorePats <GLOBAL_STORE_DWORDX2, atomic_store_64_global, v4i16>;
----------------
harrisonGPU wrote:

I have supported lds and added lit tests. Thanks.

https://github.com/llvm/llvm-project/pull/192904