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

[Jay Foad via cfe-commits]

================
@@ -402,34 +413,30 @@ Value *AMDGPUAtomicOptimizerImpl::buildReduction(IRBuilder<> &B,
 
   // Reduce within each pair of rows (i.e. 32 lanes).
   assert(ST->hasPermLaneX16());
-  V = B.CreateBitCast(V, IntNTy);
----------------
jayfoad wrote:

Please submit an NFC cleanup patch that just removes unnecessary bitcasting, before adding support for new atomic operations.

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