[inductor] Use runtime estimations in iterative reorder collectives pass (pytorch#167080)

Split of pytorch#162469 to be under 2K
reorder iterative part

Pull Request resolved: pytorch#167080
Approved by: https://github.com/eellison

Author: IvanKobzarev

test/distributed/test_inductor_collectives.py

@@ -1985,6 +1985,7 @@ def _reorder_communication_preserving_peak_memory(
                     "bucket_reduce_scatters_fx_bucket_size_determinator": lambda _: 2,
                     "reorder_for_compute_comm_overlap": True,
                     "reorder_for_compute_comm_overlap_passes": [
+                        _reorder_communication_preserving_peak_memory,
                         sink_waits_iterative,
                         _reorder_communication_preserving_peak_memory,
                     ],
@@ -2046,11 +2047,6 @@ def _reorder_communication_preserving_peak_memory(
         assert node_stats is not None
         self.assertTrue(isinstance(node_stats, dict))
         self.assertEqual(len(node_stats), 4)
-        it = iter(node_stats.values())
-        node_stat0 = next(it)
-        self.assertTrue(node_stat0.limiting_factor == "None")
-        node_stat1 = next(it)
-        self.assertTrue("collective ordering" in node_stat1.limiting_factor)
 
     @skipIfXpu  # https://github.com/intel/torch-xpu-ops/issues/1581
     @unittest.skipIf(not HAS_GPU, "Inductor+gpu needs triton and recent GPU arch")