[llvm] [X86] Lowering of load atomic float via cast (PR #117189)

[Matt Arsenault via llvm-commits]

================
@@ -137,6 +137,10 @@ static cl::opt<bool> VerifyNoAliasScopeDomination(
     cl::desc("Ensure that llvm.experimental.noalias.scope.decl for identical "
              "scopes are not dominating"));
 
+static cl::opt<bool> AllowAtomicVector(
----------------
arsenm wrote:

This patch should not touch the verifier, and there should never be an option to make IR valid or not 

https://github.com/llvm/llvm-project/pull/117189