Fix overzealous concretization root->rfactor propagation

csrc/dynamic_transform.cpp

@@ -542,6 +542,9 @@ void DynamicTransformConcretizer::mutate(TensorView* tv) {
 
       // Update the IterType of each output
       for (auto out_id : ir_utils::filterByType<IterDomain>(expr->outputs())) {
+        if (!out_id->isSymbolic()) {
+          continue;
+        }
         auto concretized_out_id =
             IterDomainBuilder(out_id).iter_type(iter_type).build();
         registerConcretization(out_id, concretized_out_id);

test/test_dynamic_transform.cpp

@@ -1001,4 +1001,35 @@ TEST_F(NVFuserTest, DynamicPadShmoo_CUDA) {
   reductionDynamicPadAddFusion(invocations);
 }
 
+// Test that a Symbolic root/Broadcast rfactor is not  concretized to
+// Iteration/Iteration
+TEST_F(NVFuserTest, FusionDynamicSliceToBroadcast_CUDA) {
+  std::unique_ptr<Fusion> fusion_ptr = std::make_unique<Fusion>();
+  Fusion& fusion = *fusion_ptr.get();
+  FusionGuard fg(fusion_ptr.get());
+  auto tv0 = makeSymbolicTensor(1);
+  fusion.addInput(tv0);
+  // tv0[:2] introduces symbolic IterDomain
+  auto tv1 = slice(
+      tv0, {{fusion.zeroVal(), IrBuilder::create<Int>(2), fusion.oneVal()}});
+  // tv1 has Broadcast rfactor, Iteration root
+  auto tv2 = slice(tv1, {{fusion.zeroVal(), fusion.oneVal(), fusion.oneVal()}});
+  // tv2 has a Symbolic root related to a Broadcast rfactor through a Resize op
+  fusion.addOutput(tv2);
+
+  // At concretization, tv1's rfactor will be set to Iteration, which will
+  // propagate to tv2s root. This test will test that when tv2 root is
+  // concretized to Iteration, it does not wind up overwriting the Broadcast
+  // rfactor.
+
+  FusionExecutorCache fusion_executor_cache(std::move(fusion_ptr));
+  auto options = at::TensorOptions().dtype(at::kFloat).device(at::kCUDA, 0);
+  at::Tensor at0 = at::randn({5}, options);
+  std::vector<c10::IValue> aten_inputs = {at0};
+  auto outputs = fusion_executor_cache.runFusionWithInputs(aten_inputs);
+  auto at1 = at::slice(at0, 0, 0, 2);
+  auto at2 = at::slice(at1, 0, 0, 1);
+  testValidate(&fusion, outputs, aten_inputs, {at2}, __LINE__, __FILE__);
+}
+
 } // namespace nvfuser