diff --git a/imperative/python/megengine/tools/load_network_and_run.py b/imperative/python/megengine/tools/load_network_and_run.py
index 2b56a2fb..ba9cd76f 100755
--- a/imperative/python/megengine/tools/load_network_and_run.py
+++ b/imperative/python/megengine/tools/load_network_and_run.py
@@ -156,6 +156,9 @@ def run_model(args, graph, inputs, outputs, data):
 
     func = graph.compile(outputs)
 
+    if args.get_static_mem_info:
+        func.get_static_memory_alloc_info(args.get_static_mem_info)
+
     def run():
         if not args.embed_input:
             for key in inp_dict:
@@ -389,6 +392,11 @@ def main():
         help="embed input data as SharedDeviceTensor in model, "
         "to remove memory copy for inputs",
     )
+    parser.add_argument(
+        "--get-static-mem-info",
+        type=str,
+        help="Record the static graph's static memory info.",
+    )
     args = parser.parse_args()
 
     if args.verbose:
diff --git a/imperative/python/src/graph_rt.cpp b/imperative/python/src/graph_rt.cpp
index 15768f7e..8bb8f5ca 100644
--- a/imperative/python/src/graph_rt.cpp
+++ b/imperative/python/src/graph_rt.cpp
@@ -215,7 +215,10 @@ void init_graph_rt(py::module m) {
                 }
             }
             return ret;
-        });
+        })
+        .def("get_static_memory_alloc_info",
+             &cg::AsyncExecutable::get_static_memory_alloc_info,
+             py::call_guard<py::gil_scoped_release>());
 
     auto PyComputingGraph = py::class_<cg::ComputingGraph, std::shared_ptr<cg::ComputingGraph>>(m, "ComputingGraph")
         .def(py::init(py::overload_cast<>(&cg::ComputingGraph::make)))