github-starred/nvitop
2
0
Fork 0
mirror of https://github.com/XuehaiPan/nvitop.git synced 2026-05-15 14:15:55 -06:00
Code Issues 29 Projects Packages Wiki Activity Actions

[GH-ISSUE #69] [Bug] Execution Freeze while using CudaDevice.count() in global scope #41

New issue
Closed
opened 2026-05-05 03:23:05 -06:00 by gitea-mirror · 5 comments
gitea-mirror commented 2026-05-05 03:23:05 -06:00
Owner
Copy link

Originally created by @aradhyamathur on GitHub (Apr 10, 2023).
Original GitHub issue: https://github.com/XuehaiPan/nvitop/issues/69

Originally assigned to: @XuehaiPan on GitHub.

Required prerequisites

  • I have read the documentation https://nvitop.readthedocs.io.
  • I have searched the Issue Tracker that this hasn't already been reported. (comment there if it has.)
  • I have tried the latest version of nvitop in a new isolated virtual environment.

Questions

Upon calling print(CudaDevice.count()) I receive the following error and the execution gets stuck and have to interrupt it manually. Can you please guide ?
nvitop version 1.1.1
NVIDIA-SMI 470.182.03 Driver Version: 470.182.03 CUDA Version: 11.4


  File "schedule_clients.py", line 27, in <module>
    print(CudaDevice.count())
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2132, in count
    return len(super().parse_cuda_visible_devices())
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 488, in parse_cuda_visible_devices
    return parse_cuda_visible_devices(cuda_visible_devices)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2357, in parse_cuda_visible_devices
    return _parse_cuda_visible_devices(cuda_visible_devices, format='index')
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2491, in _parse_cuda_visible_devices
    raw_uuids = _parse_cuda_visible_devices_to_uuids(cuda_visible_devices, verbose=False)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2616, in _parse_cuda_visible_devices_to_uuids
    parser.start()
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/process.py", line 121, in start
    self._popen = self._Popen(self)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/context.py", line 284, in _Popen
    return Popen(process_obj)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__
    super().__init__(process_obj)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__
    self._launch(process_obj)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 42, in _launch
    prep_data = spawn.get_preparation_data(process_obj._name)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/spawn.py", line 154, in get_preparation_data
    _check_not_importing_main()
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/spawn.py", line 134, in _check_not_importing_main
    raise RuntimeError('''
RuntimeError: 
        An attempt has been made to start a new process before the
        current process has finished its bootstrapping phase.

        This probably means that you are not using fork to start your
        child processes and you have forgotten to use the proper idiom
        in the main module:

            if __name__ == '__main__':
                freeze_support()
                ...

        The "freeze_support()" line can be omitted if the program
        is not going to be frozen to produce an executable.
^CTraceback (most recent call last):
  File "/home/aradhya/stable-dreamfusionBkp/schedule_clients.py", line 27, in <module>
    print(CudaDevice.count())
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2132, in count
    return len(super().parse_cuda_visible_devices())
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 488, in parse_cuda_visible_devices
    return parse_cuda_visible_devices(cuda_visible_devices)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2357, in parse_cuda_visible_devices
    return _parse_cuda_visible_devices(cuda_visible_devices, format='index')
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2491, in _parse_cuda_visible_devices
    raw_uuids = _parse_cuda_visible_devices_to_uuids(cuda_visible_devices, verbose=False)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2623, in _parse_cuda_visible_devices_to_uuids
    result = queue.get()
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/queues.py", line 365, in get
    res = self._reader.recv_bytes()
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 216, in recv_bytes
    buf = self._recv_bytes(maxlength)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 414, in _recv_bytes
    buf = self._recv(4)
  File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 379, in _recv
    chunk = read(handle, remaining)
KeyboardInterrupt
Originally created by @aradhyamathur on GitHub (Apr 10, 2023). Original GitHub issue: https://github.com/XuehaiPan/nvitop/issues/69 Originally assigned to: @XuehaiPan on GitHub. ### Required prerequisites - [X] I have read the documentation <https://nvitop.readthedocs.io>. - [X] I have searched the [Issue Tracker](https://github.com/XuehaiPan/nvitop/issues) that this hasn't already been reported. (comment there if it has.) - [X] I have tried the latest version of nvitop in a new isolated virtual environment. ### Questions Upon calling `print(CudaDevice.count())` I receive the following error and the execution gets stuck and have to interrupt it manually. Can you please guide ? nvitop version 1.1.1 NVIDIA-SMI 470.182.03 Driver Version: 470.182.03 CUDA Version: 11.4 ``` File "schedule_clients.py", line 27, in <module> print(CudaDevice.count()) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2132, in count return len(super().parse_cuda_visible_devices()) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 488, in parse_cuda_visible_devices return parse_cuda_visible_devices(cuda_visible_devices) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2357, in parse_cuda_visible_devices return _parse_cuda_visible_devices(cuda_visible_devices, format='index') File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2491, in _parse_cuda_visible_devices raw_uuids = _parse_cuda_visible_devices_to_uuids(cuda_visible_devices, verbose=False) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2616, in _parse_cuda_visible_devices_to_uuids parser.start() File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/process.py", line 121, in start self._popen = self._Popen(self) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/context.py", line 284, in _Popen return Popen(process_obj) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__ super().__init__(process_obj) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__ self._launch(process_obj) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 42, in _launch prep_data = spawn.get_preparation_data(process_obj._name) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/spawn.py", line 154, in get_preparation_data _check_not_importing_main() File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/spawn.py", line 134, in _check_not_importing_main raise RuntimeError(''' RuntimeError: An attempt has been made to start a new process before the current process has finished its bootstrapping phase. This probably means that you are not using fork to start your child processes and you have forgotten to use the proper idiom in the main module: if __name__ == '__main__': freeze_support() ... The "freeze_support()" line can be omitted if the program is not going to be frozen to produce an executable. ^CTraceback (most recent call last): File "/home/aradhya/stable-dreamfusionBkp/schedule_clients.py", line 27, in <module> print(CudaDevice.count()) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2132, in count return len(super().parse_cuda_visible_devices()) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 488, in parse_cuda_visible_devices return parse_cuda_visible_devices(cuda_visible_devices) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2357, in parse_cuda_visible_devices return _parse_cuda_visible_devices(cuda_visible_devices, format='index') File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2491, in _parse_cuda_visible_devices raw_uuids = _parse_cuda_visible_devices_to_uuids(cuda_visible_devices, verbose=False) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/site-packages/nvitop/api/device.py", line 2623, in _parse_cuda_visible_devices_to_uuids result = queue.get() File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/queues.py", line 365, in get res = self._reader.recv_bytes() File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 216, in recv_bytes buf = self._recv_bytes(maxlength) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 414, in _recv_bytes buf = self._recv(4) File "/home/aradhya/anaconda3/envs/dreamfuse/lib/python3.9/multiprocessing/connection.py", line 379, in _recv chunk = read(handle, remaining) KeyboardInterrupt ```
gitea-mirror 2026-05-05 03:23:05 -06:00
gitea-mirror commented 2026-05-05 03:23:09 -06:00
Author
Owner
Copy link

@aradhyamathur commented on GitHub (Apr 10, 2023):

Seems like it pertains to name, i was directly running without setting the namespace, which led to the above error in the script, whereas in the ipython shell it is already set by already.

<!-- gh-comment-id:1501608155 --> @aradhyamathur commented on GitHub (Apr 10, 2023): Seems like it pertains to __name__, i was directly running without setting the namespace, which led to the above error in the script, whereas in the ipython shell it is already set by already.
gitea-mirror commented 2026-05-05 03:23:10 -06:00
Author
Owner
Copy link

@XuehaiPan commented on GitHub (Apr 10, 2023):

Hi @aradhyamathur, thanks for raising this. Having CudaDevice.count() in global scope is a common use case. I think this should be considered a bug. Executing the CUDA_VISIBLE_DEVICE parser in the global scope will always reports this annoying message:

# test.py

from nvitop import CudaDevice

print(CudaDevice.count())

RuntimeError: 
        An attempt has been made to start a new process before the
        current process has finished its bootstrapping phase.

        This probably means that you are not using fork to start your
        child processes and you have forgotten to use the proper idiom
        in the main module:

            if __name__ == '__main__':
                freeze_support()
                ...

        The "freeze_support()" line can be omitted if the program
        is not going to be frozen to produce an executable.

I opened a PR #70 to resolve this. Could you try the patch with the following command?

pip3 install git+https://github.com/XuehaiPan/nvitop@spawn-subprocess
<!-- gh-comment-id:1501790017 --> @XuehaiPan commented on GitHub (Apr 10, 2023): Hi @aradhyamathur, thanks for raising this. Having `CudaDevice.count()` in global scope is a common use case. I think this should be considered a bug. Executing the `CUDA_VISIBLE_DEVICE` parser in the global scope will always reports this annoying message: ```python # test.py from nvitop import CudaDevice print(CudaDevice.count()) ``` ```python RuntimeError: An attempt has been made to start a new process before the current process has finished its bootstrapping phase. This probably means that you are not using fork to start your child processes and you have forgotten to use the proper idiom in the main module: if __name__ == '__main__': freeze_support() ... The "freeze_support()" line can be omitted if the program is not going to be frozen to produce an executable. ``` I opened a PR #70 to resolve this. Could you try the patch with the following command? ```bash pip3 install git+https://github.com/XuehaiPan/nvitop@spawn-subprocess ```
gitea-mirror commented 2026-05-05 03:23:11 -06:00
Author
Owner
Copy link

@XuehaiPan commented on GitHub (Apr 11, 2023):

This fix is included in release 1.1.2.

pip3 install --upgrade nvitop
<!-- gh-comment-id:1503375371 --> @XuehaiPan commented on GitHub (Apr 11, 2023): This fix is included in release 1.1.2. ```bash pip3 install --upgrade nvitop ```
gitea-mirror commented 2026-05-05 03:23:11 -06:00
Author
Owner
Copy link

@aradhyamathur commented on GitHub (Apr 11, 2023):

Thanks I'll check.

<!-- gh-comment-id:1503719084 --> @aradhyamathur commented on GitHub (Apr 11, 2023): Thanks I'll check.
gitea-mirror commented 2026-05-05 03:23:12 -06:00
Author
Owner
Copy link

@aradhyamathur commented on GitHub (Apr 11, 2023):

Yeah @XuehaiPan it works, thanks.

<!-- gh-comment-id:1503724852 --> @aradhyamathur commented on GitHub (Apr 11, 2023): Yeah @XuehaiPan it works, thanks.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
drop-windows-curses
pep-639
v1.6.2
v1.6.1
v1.6.0
v1.5.3
v1.5.2
v1.5.1
v1.5.0
v1.4.2
v1.4.1
v1.4.0
v1.3.2
v1.3.1
v1.3.0
v1.2.0
v1.1.2
v1.1.1
v1.1.0
v1.0.0
v0.11.0
v0.10.2
v0.10.1
v0.10.0
v0.9.0
v0.8.1
v0.8.0
v0.7.3
v0.7.2
v0.7.1
v0.7.0
v0.6.2
v0.6.1
v0.6.0
v0.5.6
v0.5.5
v0.5.4
v0.5.3
v0.5.2.2
v0.5.2.1
v0.5.2
v0.5.1.4
v0.5.1.3
v0.5.1.2
v0.5.1.1
v0.5.1
v0.5.0
v0.4.4.1
v0.4.4
v0.4.3
v0.4.2.2
v0.4.2.1
v0.4.2
v0.4.1
v0.4.0
v0.3.6.2
v0.3.6.1
v0.3.6
v0.3.5.6
v0.3.5.5
v0.3.5.4
v0.3.5.3
v0.3.5.2
v0.3.5.1
v0.3.5
v0.3.4.1
v0.3.4
v0.3.3
v0.3.2.1
v0.3.2
v0.3.1
v0.3.0.2
v0.3.0.1
v0.3.0
v0.2.5.2
v0.2.5.1
v0.2.5
v0.2.4.3
v0.2.4.2
v0.2.4.1
v0.2.4
v0.2.3
v0.2.2
v0.2.1
v0.2
v0.1
Labels
Clear labels   
api

   
bug

   
bug

   
cli / tui

   
dependencies

   
documentation

   
documentation

   
documentation

   
duplicate

   
enhancement

   
exporter

   
invalid

   
pull-request

Mirrored from GitHub Pull Request

  
pynvml

   
question

   
question

   
upstream

   
wontfix
No labels
api
bug
bug
cli / tui
dependencies
documentation
documentation
documentation
duplicate
enhancement
exporter
invalid
pull-request
pynvml
question
question
upstream
wontfix
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/nvitop#41
No description provided.
Delete branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?