[clang] [llvm] [AMDGPU] Enable atomic optimizer for 64 bit divergent values (PR #96473)

[Matt Arsenault via cfe-commits]

================
@@ -178,6 +178,20 @@ bool AMDGPUAtomicOptimizerImpl::run(Function &F) {
   return Changed;
 }
 
+static bool shouldOptimize(Type *Ty) {
----------------
arsenm wrote:

Better name that expresses why this type is handleable. 

Also in a follow up, really should cover the i16/half/bfloat and 2 x half, 2 x bfloat cases 

https://github.com/llvm/llvm-project/pull/96473