diff --git a/tests/end2end/test_end2end_cnv_w1a1.py b/tests/end2end/test_end2end_cnv_w1a1.py
index c3359dcc82650bf0e9e8a5bc5276f5ca770ee96c..688e1e55c2851841d0c369e326ac954c507c3e11 100644
--- a/tests/end2end/test_end2end_cnv_w1a1.py
+++ b/tests/end2end/test_end2end_cnv_w1a1.py
@@ -72,6 +72,7 @@ from finn.util.test import get_test_model_trained, load_test_checkpoint_or_skip
 from finn.transformation.fpgadataflow.annotate_resources import AnnotateResources
 from finn.transformation.fpgadataflow.prepare_rtlsim import PrepareRTLSim
 from finn.transformation.fpgadataflow.insert_fifo import InsertFIFO
+from finn.core.throughput_test import throughput_test_rtlsim
 
 build_dir = "/tmp/" + os.environ["FINN_INST_NAME"]
 test_pynq_board = os.getenv("PYNQ_BOARD", default="Pynq-Z1")
@@ -221,6 +222,20 @@ def test_end2end_cnv_w1a1_verify_dataflow_part():
     assert np.isclose(res_cppsim, res_rtlsim_whole).all()
 
 
+@pytest.mark.vivado
+def test_end2end_cnv_w1a1_throughput_test_rtlsim():
+    model = load_test_checkpoint_or_skip(
+        build_dir + "/end2end_cnv_w1a1_ipstitch_whole_rtlsim.onnx"
+    )
+    model.set_metadata_prop("rtlsim_trace", "rtlsim_trace.vcd")
+    # os.environ["RTLSIM_TRACE_DEPTH"] = "4"
+    # run through IP-stitched rtlsim with increasing batch sizes and
+    # check the number of cycles it takes to execute
+    ret = throughput_test_rtlsim(model, 10)
+    # TODO check for expected performance
+    assert ret > 0
+
+
 @pytest.mark.vivado
 def test_end2end_cnv_w1a1_verify_all():
     # use the streamlined model as the "golden" model for right answers