[Mlir-commits] [mlir] [NvGpu Dialect] add rcp approxe op (PR #100965)

[Guray Ozen]

================
@@ -1339,3 +1339,19 @@ module attributes {transform.with_named_sequence} {
     transform.yield
   }
 }
+
+// CHECK-LABEL: @rcp_approx_ftz_f32
+// CHECK-SAME:  %[[IN:.*]]: vector<32x16xf32>
+func.func @rcp_approx_ftz_f32(
----------------
grypp wrote:

Would be great if you also add a failing test, so we check the verifier 

https://github.com/llvm/llvm-project/pull/100965