[clang] [Clang][AMDGPU] Add builtins for instrinsic `llvm.amdgcn.raw.buffer.store` (PR #94576)

Shilei Tian via cfe-commits cfe-commits at lists.llvm.org
Fri Jun 21 03:13:32 PDT 2024


================
@@ -149,6 +149,19 @@ BUILTIN(__builtin_amdgcn_mqsad_pk_u16_u8, "WUiWUiUiWUi", "nc")
 BUILTIN(__builtin_amdgcn_mqsad_u32_u8, "V4UiWUiUiV4Ui", "nc")
 
 BUILTIN(__builtin_amdgcn_make_buffer_rsrc, "Qbv*sii", "nc")
+BUILTIN(__builtin_amdgcn_raw_ptr_buffer_store_i8, "vcQbiiIi", "n")
----------------
shiltian wrote:

> Probably should drop the _ptr part of the name. This was more of a legacy issue in the intrinsic case, since the resource-as-vector case already took the name.

Sure, can do.

> Also not sure if we should follow the naming convention of the instruction instead (probably the gfx12 one?).

you mean things like `buffer_store_b8`? In this way, we don't need to differentiate types, just bit size.

https://github.com/llvm/llvm-project/pull/94576


More information about the cfe-commits mailing list