[clang] [WIP][AMDGPU] Support for type inferring image load/store builtins for AMDGPU (PR #140210)

[Matt Arsenault via cfe-commits]

================
@@ -635,5 +635,66 @@ TARGET_BUILTIN(__builtin_amdgcn_bitop3_b16, "ssssIUi", "nc", "bitop3-insts")
 TARGET_BUILTIN(__builtin_amdgcn_cvt_sr_bf16_f32, "V2yV2yfUiIb", "nc", "f32-to-f16bf16-cvt-sr-insts")
 TARGET_BUILTIN(__builtin_amdgcn_cvt_sr_f16_f32, "V2hV2hfUiIb", "nc", "f32-to-f16bf16-cvt-sr-insts")
 
+//===----------------------------------------------------------------------===//
+// Image load/store builtins
+//===----------------------------------------------------------------------===//
+TARGET_BUILTIN(__builtin_amdgcn_image_load_1d_v4f32_i32, "V4fiiV8iii", "nc", "")
+TARGET_BUILTIN(__builtin_amdgcn_image_load_1d_v4f16_i32, "V4hiiV8iii", "nc", "")
+TARGET_BUILTIN(__builtin_amdgcn_image_load_1darray_v4f32_i32, "V4fiiiV8iii", "nc", "")
----------------
arsenm wrote:

Can we stick with just the v4 forms for now and rely on the shuffle optimization? 

https://github.com/llvm/llvm-project/pull/140210