[Mlir-commits] [mlir] [mlir][ArmSME] Add tile slice to vector intrinsics (PR #66910)

[Cullen Rhodes]

================
@@ -334,3 +334,137 @@ llvm.func @arm_sme_vector_to_tile_vert(%tileslice : i32,
       (i32, i32, vector<[2]xi1>, vector<[2]xf64>) -> ()
   llvm.return
 }
+
+// -----
+
+llvm.func @prevent_dce.nxv16i8(vector<[16]xi8>)
----------------
c-rhodes wrote:

do we need to prevent DCE for intrinsics? We get away without it for all the other ones, I should think we do for these too? 

https://github.com/llvm/llvm-project/pull/66910