Could not find compiler for platform Host: NOT_FOUND: could not find registered compiler for platform Host -- was support for that platform linked in? [Op:__inference_call_72951]

[Charlie-Su-55]

Hello, I'm a beginner of Sionna. When I try to run the example "Realistic Multiuser MIMO OFDM Simulations" in Linux, there exist some errors. I have checked that the package sionna, CUDA, cuDNN, tensorflow etc are all installed. I wonder, if you have some similar error like this:

Traceback (most recent call last):
  File "~/venv/MU_MIMO_with_precoding.py", line 479, in <module>
    ber, bler = sim_ber(model,
                ^^^^^^^^^^^^^^
  File "~/anaconda3/envs/venv/lib/python3.11/site-packages/sionna/utils/misc.py", line 633, in sim_ber
    outputs = mc_fun(batch_size=batch_size, ebno_db=ebno_dbs[i])
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "~/anaconda3/envs/venv/lib/python3.11/site-packages/keras/utils/traceback_utils.py", line 70, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "~/anaconda3/envs/venv/lib/python3.11/site-packages/tensorflow/python/eager/execute.py", line 52, in quick_execute
    tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
tensorflow.python.framework.errors_impl.UnimplementedError: Exception encountered when calling layer 'model' (type Model).

Could not find compiler for platform Host: NOT_FOUND: could not find registered compiler for platform Host -- was support for that platform linked in? [Op:__inference_call_81905]

Call arguments received by layer 'model' (type Model):
  • batch_size=tf.Tensor(shape=(), dtype=int32)
  • ebno_db=tf.Tensor(shape=(), dtype=float32)

[SebastianCa]

Hi,
this sounds like an issue with your local TensorFlow installation. Could you please let us know your version of Sionna, Tensorflow and Python? And please try to run any other example notebook to see if this is a general issue with your setup.

[Charlie-Su-55]

By the way, the Num GPUs Available still shows 0.

[SebastianCa]

I assume you have an available GPU in your system? Can you please verify with nvidia-smi that the GPU is correctly installed?
It seems like TF does not find your GPU and there is something wrong with your system setup. Please check the following TF installation guide.

Note that the hello_world.ipynb notebook does not contain graph/XLA execution. Can you please try to run the Discover Sionna Example ?

[Charlie-Su-55]

Hello Sebastian, thanks a lot for your suggestion, now I manage to let the TF and GPU finding each other. And the 'Discover Sionna Example' can also run from begin to the end. But when I return to the Realistic Multiuser MIMO OFDM Simulations. The usage of GPU Memory reach to 23020MiB and stop after a while with errors:

---------------------------------------------------------------------------
InvalidArgumentError                      Traceback (most recent call last)
Cell In[14], line 18
     13 for perfect_csi in SIMS["perfect_csi"]:
     15     model = Model(scenario=scenario,
     16                   perfect_csi=perfect_csi)
---> 18     ber, bler = sim_ber(model,
     19                         SIMS["ebno_db"],
     20                         batch_size=128,
     21                         max_mc_iter=1000,
     22                         num_target_block_errors=1000)
     24     SIMS["ber"].append(list(ber.numpy()))
     25     SIMS["bler"].append(list(bler.numpy()))

File ~/lib/python3.10/site-packages/sionna/utils/misc.py:633, in sim_ber(mc_fun, ebno_dbs, batch_size, max_mc_iter, soft_estimates, num_target_bit_errors, num_target_block_errors, early_stop, graph_mode, verbose, forward_keyboard_interrupt, dtype)
    629 for ii in tf.range(max_mc_iter):
    631     iter_count += 1
--> 633     outputs = mc_fun(batch_size=batch_size, ebno_db=ebno_dbs[i])
    635     # assume first and second return value is b and b_hat
    636     # other returns are ignored
    637     b = outputs[0]

File ~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py:70, in filter_traceback.<locals>.error_handler(*args, **kwargs)
     67     filtered_tb = _process_traceback_frames(e.__traceback__)
     68     # To get the full stack trace, call:
     69     # `tf.debugging.disable_traceback_filtering()`
---> 70     raise e.with_traceback(filtered_tb) from None
     71 finally:
     72     del filtered_tb

File ~/lib/python3.10/site-packages/tensorflow/python/eager/execute.py:60, in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
     53   # Convert any objects of type core_types.Tensor to Tensor.
     54   inputs = [
     55       tensor_conversion_registry.convert(t)
     56       if isinstance(t, core_types.Tensor)
     57       else t
     58       for t in inputs
     59   ]
---> 60   tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
     61                                       inputs, attrs, num_outputs)
     62 except core._NotOkStatusException as e:
     63   if name is not None:

InvalidArgumentError: Exception encountered when calling layer 'model' (type Model).

Detected unsupported operations when trying to compile graph ldpc5g_decoder_1_while_body_81240_const_0[] on XLA_GPU_JIT: RaggedRange (No registered 'RaggedRange' OpKernel for XLA_GPU_JIT devices compatible with node {{node ldpc5g_decoder_1/while/RaggedWhere/RaggedRange}}){{node ldpc5g_decoder_1/while/RaggedWhere/RaggedRange}}
The op is created at: 
File "~/lib/python3.10/runpy.py", line 196, in _run_module_as_main
File "~/lib/python3.10/runpy.py", line 86, in _run_code
File "~/python3.10/site-packages/ipykernel_launcher.py", line 17, in <module>
File "~/lib/python3.10/site-packages/traitlets/config/application.py", line 1053, in launch_instance
File "~/lib/python3.10/site-packages/ipykernel/kernelapp.py", line 737, in start
File "~/lib/python3.10/site-packages/tornado/platform/asyncio.py", line 195, in start
File "~/lib/python3.10/asyncio/base_events.py", line 595, in run_forever
File "~/lib/python3.10/asyncio/base_events.py", line 1881, in _run_once
File "~/lib/python3.10/asyncio/events.py", line 80, in _run
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 524, in dispatch_queue
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 513, in process_one
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 418, in dispatch_shell
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 758, in execute_request
File "~/lib/python3.10/site-packages/ipykernel/ipkernel.py", line 426, in do_execute
File "~/lib/python3.10/site-packages/ipykernel/zmqshell.py", line 549, in run_cell
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3046, in run_cell
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3101, in _run_cell
File "~/lib/python3.10/site-packages/IPython/core/async_helpers.py", line 129, in _pseudo_sync_runner
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3306, in run_cell_async
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3488, in run_ast_nodes
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3548, in run_code
File "tmp/ipykernel_690781/3411179214.py", line 18, in <module>
File "~/lib/python3.10/site-packages/sionna/utils/misc.py", line 633, in sim_ber
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/training.py", line 589, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/base_layer.py", line 1149, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 96, in error_handler
File "tmp/ipykernel_690781/794116499.py", line 137, in call
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/base_layer.py", line 1149, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 96, in error_handler
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1382, in call
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1383, in call
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1002, in call
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 980, in dec_iter
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 654, in _cn_update_phi
	tf2xla conversion failed while converting ldpc5g_decoder_1_while_body_81240_const_0[]. Run with TF_DUMP_GRAPH_PREFIX=/path/to/dump/dir and --vmodule=xla_compiler=2 to obtain a dump of the compiled functions.

Stack trace for op definition: 
File "~/lib/python3.10/runpy.py", line 196, in _run_module_as_main
File "~/lib/python3.10/runpy.py", line 86, in _run_code
File "~/lib/python3.10/site-packages/ipykernel_launcher.py", line 17, in <module>
File "~/lib/python3.10/site-packages/traitlets/config/application.py", line 1053, in launch_instance
File "~/lib/python3.10/site-packages/ipykernel/kernelapp.py", line 737, in start
File "~/lib/python3.10/site-packages/tornado/platform/asyncio.py", line 195, in start
File "~/lib/python3.10/asyncio/base_events.py", line 595, in run_forever
File "~/lib/python3.10/asyncio/base_events.py", line 1881, in _run_once
File "~/lib/python3.10/asyncio/events.py", line 80, in _run
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 524, in dispatch_queue
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 513, in process_one
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 418, in dispatch_shell
File "~/lib/python3.10/site-packages/ipykernel/kernelbase.py", line 758, in execute_request
File "~/lib/python3.10/site-packages/ipykernel/ipkernel.py", line 426, in do_execute
File "~/lib/python3.10/site-packages/ipykernel/zmqshell.py", line 549, in run_cell
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3046, in run_cell
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3101, in _run_cell
File "~/lib/python3.10/site-packages/IPython/core/async_helpers.py", line 129, in _pseudo_sync_runner
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3306, in run_cell_async
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3488, in run_ast_nodes
File "~/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3548, in run_code
File "tmp/ipykernel_690781/3411179214.py", line 18, in <module>
File "~/lib/python3.10/site-packages/sionna/utils/misc.py", line 633, in sim_ber
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/training.py", line 589, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/base_layer.py", line 1149, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 96, in error_handler
File "tmp/ipykernel_690781/794116499.py", line 137, in call
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 65, in error_handler
File "~/lib/python3.10/site-packages/keras/src/engine/base_layer.py", line 1149, in __call__
File "~/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 96, in error_handler
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1382, in call
File "~/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1383, in call
File "~/lib/python3.10/site-packages/sionna/fec/ldpc/decoding.py", line 1002, in call

	 [[ldpc5g_decoder_1/while]]
	tf2xla conversion failed while converting __inference_call_82794[_XlaMustCompile=true,config_proto=8589078909834744431,executor_type=11160318154034397263]. Run with TF_DUMP_GRAPH_PREFIX=/path/to/dump/dir and --vmodule=xla_compiler=2 to obtain a dump of the compiled functions. [Op:__inference_call_82794]

Call arguments received by layer 'model' (type Model):
  • batch_size=tf.Tensor(shape=(), dtype=int32)
  • ebno_db=tf.Tensor(shape=(), dtype=float32)


----------------------------------------------------------------

I wonder if there would be some problem with my XLA setting or dump dictionary setting?
Thanks again for your help for these days.

[SebastianCa]

Great to see that it works now.

You are now observing a different bug related to an incompatibility with TF 2.14 (see e.g., #255).
A current workaround is to downgrade to TF 2.13 or avoid XLA mode.

[Charlie-Su-55]

(:з」∠)
Thank you soooooooooooooooooooooooo much Sebastian! The code finally run after i turn tensorflow 2.14 to 2.13!
Very kind of you, you make my day!