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

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,138 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+
----------------
arsenm wrote:

Better name would be scalar-load instead of sreg 

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