cherry pick 3660 to release/2.8 (#3661)

Author: lanluo-nvidia

.github/workflows/build-test-windows.yml

@@ -230,7 +230,7 @@ jobs:
         cd dynamo
         python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
         python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
+        ../../../packaging/vc_env_helper.bat python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
         popd
 
   tests-py-dynamo-core:

py/torch_tensorrt/dynamo/lowering/passes/constant_folding.py

@@ -103,10 +103,14 @@ def __init__(self, *args: Any, **kwargs: Any) -> None:
         self.quantization_ops: Set[torch._ops.OpOverload] = set()
         try:
             # modelopt import ensures torch.ops.tensorrt.quantize_op.default is registered
-            import modelopt.torch.quantization as mtq
+            import modelopt.torch.quantization as mtq  # noqa: F401
 
             assert torch.ops.tensorrt.quantize_op.default
+            assert torch.ops.tensorrt.dynamic_block_quantize_op.default
             self.quantization_ops.add(torch.ops.tensorrt.quantize_op.default)
+            self.quantization_ops.add(
+                torch.ops.tensorrt.dynamic_block_quantize_op.default
+            )
         except Exception as e:
             pass