brmarkus
commented
1 month ago I just tried to reproduce the Juypter notebook "notebooks/3D-segmentation-point-clouds/3D-segmentation-point-clouds.ipynb". On a MS-Win 11 Laptop (Intel Core Ultra 7 155H). With a clean Python venv, using current HEAD-revision of this notebooks repo. I can confirm the inference is working when using the CPU. But it fails when using the GPU. (NPU not working due to use of dynamic shapes)
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
Cell In[14], line 2
1 # Inference
----> 2 compiled_model = core.compile_model(model=model, device_name=device.value)
3 output_layer = compiled_model.output(0)
4 result = compiled_model([point])[output_layer]
File [C:\localdisk\OpenVINO-MSWin\openvino_env\Lib\site-packages\openvino\runtime\ie_api.py:543](file:///C:/localdisk/OpenVINO-MSWin/openvino_env/Lib/site-packages/openvino/runtime/ie_api.py#line=542), in Core.compile_model(self, model, device_name, config, weights)
538 if device_name is None:
539 return CompiledModel(
540 super().compile_model(model, {} if config is None else config),
541 )
542 return CompiledModel(
--> 543 super().compile_model(model, device_name, {} if config is None else config),
544 )
545 else:
546 if device_name is None:
RuntimeError: Exception from src\inference\src\cpp\core.cpp:107:
Exception from src\inference\src\dev\plugin.cpp:53:
Check 'false' failed at src\plugins\intel_gpu\src\plugin\program_builder.cpp:185:
[GPU] ProgramBuilder build failed!
Program build failed(0_part_11):
21:5264:5: error: token is not a valid binary operator in a preprocessor subexpression
#if INPUT_LEFTOVERS
^~~~~~~~~~~~~~~
21:5133:38: note: expanded from macro 'INPUT_LEFTOVERS'
....
21:7301:5: error: token is not a valid binary operator in a preprocessor subexpression
#if INPUT_LEFTOVERS
^~~~~~~~~~~~~~~
21:7170:38: note: expanded from macro 'INPUT_LEFTOVERS'
....
21:9338:5: error: token is not a valid binary operator in a preprocessor subexpression
#if INPUT_LEFTOVERS
^~~~~~~~~~~~~~~
21:9207:38: note: expanded from macro 'INPUT_LEFTOVERS'
....According to a call to %pip freeze I got OpenVINO installed in version openvino==2024.3.0... your reported problem seems to exist with older versions already.
YOODS-Xu
Describe the bug openVINO2024.3.0 gave RuntimeError:[GPU] ProgramBuilder build failed! at notebook:3D-segmentation-point-clouds, when using Integrated GPU both 11 Gen corei7 and 13 Gen corei9. openVINO2023.3.0 worked well but integrated GPU were slower than CPUs.
Expected behavior Did any other's openVINO2024.3.0 work well at notebook:3D-segmentation-point-clouds, when using Integrated GPU? What is the matter with me, or is it dispossible? How could my Integrated GPU be faster than CPU using segmetation models such as 3D-segmentation-point-clouds and Mask rcnn?
Screenshots
Installation instructions (Please mark the checkbox) [yes] I followed the installation guide at https://github.com/openvinotoolkit/openvino_notebooks#-installation-guide to install the notebooks.
Environment information Only next is not ok, but I donot know how to fix it, and if it mattered. Jupyter kernel installed for openvino_env:
print(f"kernel spec={KernelSpecManager().get_all_specs()}") gave as below. kernel spec={'python3': {'resource_dir': '/home2/u22-ros/venvs/openvino_env/share/jupyter/kernels/python3', 'spec': {'argv': ['python', '-m', 'ipykernel_launcher', '-f', '{connection_file}'], 'env': {}, 'display_name': 'Python 3 (ipykernel)', 'language': 'python', 'interrupt_mode': 'signal', 'metadata': {'debugger': True}}}}
After modified kernel = KernelSpecManager().get_kernel_spec("openvino_env") to kernel = KernelSpecManager().get_kernel_spec("python3") then below are not ok. Jupyter kernel Python and OpenVINO environment Python match: NOT OK
Additional context Add any other context about the problem here.