[llvm] [AMDGPU][DAG] Enable ganging up of memcpy loads/stores for AMDGPU (PR #96185)

[Matt Arsenault via llvm-commits]

================
@@ -67,6 +67,9 @@ AMDGPUTargetLowering::AMDGPUTargetLowering(const TargetMachine &TM,
   MaxStoresPerMemcpy = MaxStoresPerMemcpyOptSize = ~0U;
   MaxStoresPerMemmove = MaxStoresPerMemmoveOptSize = ~0U;
 
+  // Enable ganging up loads and stores in the memcpy DAG lowering.
----------------
arsenm wrote:

Turning it on first is fine. We should also expand the test coverage for different sizes, alignments and address space combinations. I'm surprised we don't have more test updates from this as-is 

https://github.com/llvm/llvm-project/pull/96185