ResourceWarning shows up when doing multigpu graphbolt example

[TristonC]

Runs the example under dgl/examples/multigpu/graphbolt, it shows a bit annoying ResourceWarning.

 python node_classification.py --gpu 0,1,2,3,4,5,6,7

16it [00:01, 12.86it/s]
Epoch 00009 | Average Loss 1.1217 | Accuracy 0.6525 | Time 14.4035
Testing...
25it [00:02, 11.18it/s]/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpnqgx4sch'>
  _warnings.warn(warn_message, ResourceWarning)
27it [00:02, 11.99it/s]
Test Accuracy 0.6272
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpns5n0f16'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmp4sbe1b2l'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpgtm6h_av'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpbebuzjfx'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpnp2bys02'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmp0mbq_2c8'>
  _warnings.warn(warn_message, ResourceWarning)
/usr/lib/python3.10/tempfile.py:1008: ResourceWarning: Implicitly cleaning up <TemporaryDirectory '/tmp/tmpcirawy77'>
  _warnings.warn(warn_message, ResourceWarning)

[mfbalin]

I will get to this issue if I can find the time. Meanwhile let me unassign myself so that if anyone else is free, they can take care of this issue.

My priority is to finish the remaining features right now.

[mfbalin]

I can't reproduce the issue on the pytorch 24-07 container.

root@a100cse:/localscratch/dgl-3/examples/graphbolt/pyg/labor# python ../../../multigpu/graphbolt/node_classification.py --gpu 0,1
Training with 2 gpus.
The dataset is already preprocessed.
[W905 21:16:53.916054495 CUDAAllocatorConfig.h:28] Warning: expandable_segments not supported on this platform (function operator())
[W905 21:16:53.918450461 CUDAAllocatorConfig.h:28] Warning: expandable_segments not supported on this platform (function operator())
Training...
96it [00:02, 37.52it/s]
Validating...
20it [00:00, 48.91it/s]
Epoch 00000 | Average Loss 1.6953 | Accuracy 0.8335 | Time 3.0184
96it [00:02, 45.63it/s]
Validating...
20it [00:00, 50.92it/s]
Epoch 00001 | Average Loss 0.7198 | Accuracy 0.8611 | Time 2.5459
96it [00:02, 45.87it/s]
Validating...
20it [00:00, 51.33it/s]
Epoch 00002 | Average Loss 0.5773 | Accuracy 0.8751 | Time 2.5282
96it [00:02, 45.91it/s]
Validating...
20it [00:00, 51.38it/s]
Epoch 00003 | Average Loss 0.5099 | Accuracy 0.8815 | Time 2.5286
96it [00:02, 46.05it/s]
Validating...
20it [00:00, 51.62it/s]
Epoch 00004 | Average Loss 0.4715 | Accuracy 0.8868 | Time 2.5248
96it [00:02, 45.91it/s]
Validating...
20it [00:00, 51.77it/s]
Epoch 00005 | Average Loss 0.4428 | Accuracy 0.8908 | Time 2.5248
96it [00:02, 45.82it/s]
Validating...
20it [00:00, 51.51it/s]
Epoch 00006 | Average Loss 0.4214 | Accuracy 0.8927 | Time 2.5373
96it [00:02, 45.92it/s]
Validating...
20it [00:00, 51.54it/s]
Epoch 00007 | Average Loss 0.4086 | Accuracy 0.8960 | Time 2.5373
96it [00:02, 45.86it/s]
Validating...
20it [00:00, 51.94it/s]
Epoch 00008 | Average Loss 0.3919 | Accuracy 0.8970 | Time 2.5347
96it [00:02, 45.57it/s]
Validating...
20it [00:00, 53.51it/s]
Epoch 00009 | Average Loss 0.3831 | Accuracy 0.8984 | Time 2.5382
Testing...
1081it [00:18, 57.49it/s]
Test Accuracy 0.7647

[mfbalin]

@TristonC Can you provide more details on your environment? I can't reproduce the issue on my machines.