[clang] [llvm] [AMDGPU] Asynchronous loads from global/buffer to LDS on pre-GFX12 (PR #180466)

[Sameer Sahasrabuddhe via cfe-commits]

================
@@ -0,0 +1,238 @@
+.. _amdgpu-async-operations:
----------------
ssahasra wrote:

I had already added it in one place. Now I added a ".async" variant of the one intrinsic that I found in the table: llvm.amdgcn.load.to.lds.p<1/7> 

The overall existing documentation in to AMDGPUsage now covers the following, along with their async versions:
-  llvm.amdgcn.load.to.lds.p<1/7>
- global_load_lds 
- buffer/global_load instructions with the lds flag 

I have not tried to deal with missing entries in the table. Frankly, this change has languished too long to delay further over documentation that was already missing.

https://github.com/llvm/llvm-project/pull/180466