Skip to content

tests.test_spatial_resample error #4432

@wyli

Description

@wyli

Describe the bug
randomly run into this issue:

[2022-06-01T14:46:36.804Z] ======================================================================
[2022-06-01T14:46:36.804Z] ERROR: test_ill_affine_1 (tests.test_spatial_resample.TestSpatialResample)
[2022-06-01T14:46:36.804Z] ----------------------------------------------------------------------
[2022-06-01T14:46:36.804Z] Traceback (most recent call last):
[2022-06-01T14:46:36.804Z]   File "/home/jenkins/agent/workspace/MONAI-premerge/monai/monai/transforms/spatial/array.py", line 274, in __call__
[2022-06-01T14:46:36.804Z]     if pytorch_after(1, 8, 0)
[2022-06-01T14:46:36.804Z] RuntimeError: CUDA error: an illegal memory access was encountered
[2022-06-01T14:46:36.804Z] CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
[2022-06-01T14:46:36.804Z] For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
[2022-06-01T14:46:36.804Z] 
[2022-06-01T14:46:36.804Z] During handling of the above exception, another exception occurred:
[2022-06-01T14:46:36.804Z] 
[2022-06-01T14:46:36.804Z] Traceback (most recent call last):
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/parameterized/parameterized.py", line 533, in standalone_func
[2022-06-01T14:46:36.804Z]     return func(*(a + p.args), **p.kwargs)
[2022-06-01T14:46:36.804Z]   File "/home/jenkins/agent/workspace/MONAI-premerge/monai/tests/test_spatial_resample.py", line 161, in test_ill_affine
[2022-06-01T14:46:36.804Z]     SpatialResample()(img=img, dst_affine=dst_affine)
[2022-06-01T14:46:36.804Z]   File "/home/jenkins/agent/workspace/MONAI-premerge/monai/monai/utils/deprecate_utils.py", line 217, in _wrapper
[2022-06-01T14:46:36.804Z]     return func(*args, **kwargs)
[2022-06-01T14:46:36.804Z]   File "/home/jenkins/agent/workspace/MONAI-premerge/monai/monai/transforms/spatial/array.py", line 278, in __call__
[2022-06-01T14:46:36.804Z]     raise ValueError(f"src affine is not invertible: {src_affine_}") from e
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor.py", line 628, in __format__
[2022-06-01T14:46:36.804Z]     return object.__format__(self, format_spec)
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor.py", line 305, in __repr__
[2022-06-01T14:46:36.804Z]     return torch._tensor_str._str(self)
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor_str.py", line 434, in _str
[2022-06-01T14:46:36.804Z]     return _str_intern(self)
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor_str.py", line 409, in _str_intern
[2022-06-01T14:46:36.804Z]     tensor_str = _tensor_str(self, indent)
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor_str.py", line 264, in _tensor_str
[2022-06-01T14:46:36.804Z]     formatter = _Formatter(get_summarized_data(self) if summarize else self)
[2022-06-01T14:46:36.804Z]   File "/usr/local/lib/python3.7/dist-packages/torch/_tensor_str.py", line 100, in __init__
[2022-06-01T14:46:36.804Z]     nonzero_finite_vals = torch.masked_select(tensor_view, torch.isfinite(tensor_view) & tensor_view.ne(0))
[2022-06-01T14:46:36.804Z] RuntimeError: CUDA error: an illegal memory access was encountered
[2022-06-01T14:46:36.804Z] CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
[2022-06-01T14:46:36.804Z] For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
[2022-06-01T14:46:36.804Z] 

xform = (
torch.linalg.solve(src_affine_, dst_affine)
if pytorch_after(1, 8, 0)
else torch.solve(dst_affine, src_affine_).solution # type: ignore
)

torch 1.10.2, cuda 10.2

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions