Add UT error cases to skip list for LNL windows device #1740
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Cheng, Penghui <[email protected]>
There was a problem hiding this comment.
Pull Request Overview
Adds new unit test error cases to the Windows LNL device skip list, preventing known failures in interpolation and neural network tests.
- Skips four failing interpolation tests in
test_ops_xpu.py - Skips two failing neural network tests in
test_nn_xpu.py
Comments suppressed due to low confidence (2)
test/xpu/skip_list_win_lnl.py:49
- The skip-list entry for
test_CTCLoss_long_targetslikely needs the_xpusuffix to match the actual test name (test_CTCLoss_long_targets_xpu). Please verify the exact test function name and update accordingly.
"test_CTCLoss_long_targets",
test/xpu/skip_list_win_lnl.py:52
- The skip-list entry for
test_no_gradmay be missing the_xpusuffix (e.g.,test_no_grad_xpu). Ensure it matches the actual test name so the failure is correctly skipped.
"test_no_grad",
Signed-off-by: Cheng, Penghui <[email protected]>
@sys_pytorchxpubot triage result for run 15628313196Triage bot UT analaysis result for reference only, please note unique error message only report once:
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "Open",
"issue_owner": "riverliuintel",
"issue_description": "The test is failing because the ceil function is not implemented for complex128 tensors on XPU. The error indicates that the operation 'foreach_unary_op_xpu_' is not supported for 'ComplexDouble', and the ceil function does not handle complex inputs on XPU devices.",
"root_causes": [
"The ceil function is not implemented for complex128 tensors on XPU.",
"Lack of support for complex number operations in the XPU backend."
],
"suggested_solutions": [
"Implement the ceil function for complex128 tensors in the XPU backend.",
"Add necessary kernels or modify existing ones to handle complex data types.",
"Ensure all relevant unary operations support complex numbers on XPU."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "Open",
"issue_owner": "riverliuintel",
"issue_description": "The test case failed with a NotImplementedError and RuntimeError indicating that 'foreach_unary_op_xpu_' is not implemented for 'ComplexFloat' and 'ceil' is not supported for complex inputs. The error occurs during a foreach operation on a list of complex64 tensors on XPU.",
"root_causes": [
"Lack of implementation for 'foreach_unary_op_xpu_' for ComplexFloat tensors.",
"Unsupported 'ceil' operation for complex inputs on XPU.",
"Missing kernel support for complex number operations in Torch-XPU."
],
"suggested_solutions": [
"Implement the 'foreach_unary_op_xpu_' function for ComplexFloat tensors.",
"Add support for 'ceil' and other mathematical operations for complex numbers on XPU.",
"Develop and integrate the necessary kernels to handle complex number operations in the XPU backend."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "closed",
"issue_owner": "PenghuiCheng",
"issue_description": "The unit test failed with an error indicating that 'foreach_unary_op_xpu' is not implemented for 'ComplexDouble' and that the ceil function does not support complex inputs on XPU devices. This issue is specific to the handling of complex numbers in the ceil function within the foreach operations on XPU.",
"root_causes": [
"The ceil function is not implemented for complex numbers on XPU devices.",
"The foreach_unary_op_xpu operation lacks support for complex types, leading to the failure in processing complex128 tensors."
],
"suggested_solutions": [
"Implement the ceil function for complex numbers in the XPU backend to support complex128 tensors.",
"Ensure that the foreach_unary_op_xpu operation is updated to handle complex types correctly, aligning it with other supported operations on XPU."
]
}
Triage bot response: {
"similar_issue_id": "test_unary_op_tensors_on_different_devices__foreach_neg_xpu_bool",
"similar_issue_state": "closed",
"issue_owner": "fengyuan14",
"issue_description": "The test `test_parity__foreach_clamp_max_fastpath_inplace_xpu_complex64` failed due to `NotImplementedError` and `RuntimeError`, indicating that `foreach_binary_op_scalar_xpu_` is not implemented for 'ComplexFloat' and 'clamp' is not supported for complex types. The error arises from operations involving complex tensors not being properly handled in the XPU backend.",
"root_causes": [
"Missing implementation of `foreach_binary_op_scalar_xpu_` for complex types.",
"Unsupported `clamp` operation for complex tensors in the XPU backend."
],
"suggested_solutions": [
"Implement `foreach_binary_op_scalar_xpu_` for complex types to handle scalar operations correctly.",
"Add support for the `clamp` function for complex tensors in the XPU backend to ensure proper operation."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "closed",
"issue_owner": "fengyuan14",
"issue_description": "The test `test_parity__foreach_floor_fastpath_inplace_xpu_complex128` failed due to a `NotImplementedError` indicating that `foreach_unary_op_xpu_` is not implemented for 'ComplexDouble' and a `RuntimeError` stating that the `floor` function is not supported for complex inputs. This issue is related to the lack of support for certain operations on complex tensors in Torch-XPU.",
"root_causes": [
"The `floor` function and other unary operations are not implemented for complex numbers on XPU.",
"The `foreach_unary_op_xpu_` function is not implemented for the 'ComplexDouble' dtype."
],
"suggested_solutions": [
"Implement the `foreach_unary_op_xpu_` function to support 'ComplexDouble' dtype.",
"Ensure that the `floor` function and other affected operations are properly supported for complex tensors on XPU."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "closed",
"issue_owner": "fengyuan14",
"issue_description": "The unit test failed with an error indicating that 'foreach_unary_op_xpu' is not implemented for 'ComplexDouble' and that the floor function is not supported for complex inputs. The error message suggests that the XPU backend lacks support for certain operations on complex numbers, similar to previous issues.",
"root_causes": [
"Lack of implementation for unary operations on complex numbers in the XPU backend.",
"Missing support for functions like floor, ceil, etc., for complex types in XPU operations."
],
"suggested_solutions": [
"Implement the missing unary operations (e.g., floor, ceil) for complex numbers in the XPU backend.",
"Ensure that these operations are correctly handled in the XPU kernels to support complex types."
]
}
Triage bot response: {
"similar_issue_id": 628,
"similar_issue_state": "resolved",
"issue_owner": "riverliuintel",
"issue_description": "The test is failing due to a `NotImplementedError` when using `torch.sign` on complex tensors on XPU. The error suggests that the operation isn't implemented for complex numbers, and `torch.sgn` should be used instead.",
"root_causes": [
"The `foreach_unary_op_xpu_` function isn't implemented for complex float tensors.",
"The test uses `torch.sign`, which doesn't support complex numbers, instead of `torch.sgn`."
],
"suggested_solutions": [
"Implement the missing `foreach_unary_op_xpu_` function for complex float tensors in the XPU backend.",
"Modify the test to use `torch.sgn` instead of `torch.sign` for complex number operations."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "closed",
"issue_owner": "PenghuiCheng",
"issue_description": "The unit test `test_parity__foreach_trunc_fastpath_inplace_xpu_complex128` is failing due to the `foreach_unary_op_xpu_` operation not being implemented for 'ComplexDouble' and the 'trunc' function not supporting complex inputs. This issue is similar to a previously reported problem where multiple operations lacked support for complex numbers on XPU devices.",
"root_causes": [
"The 'trunc' operation is not implemented for complex numbers on XPU.",
"Missing implementation of 'foreach_unary_op_xpu_' for ComplexDouble tensors."
],
"suggested_solutions": [
"Implement the 'trunc' operation for ComplexDouble tensors on XPU.",
"Add the missing 'foreach_unary_op_xpu_' kernel for complex number support."
]
}
Triage bot response: {
"similar_issue_id": 628,
"similar_issue_state": "closed",
"issue_owner": "PenghuiCheng",
"issue_description": "The test `test_parity__foreach_trunc_fastpath_inplace_xpu_complex64` failed due to `foreach_unary_op_xpu_` not being implemented for 'ComplexFloat' and 'trunc' not supporting complex inputs. This indicates missing kernel implementations for complex64 tensors on XPU.",
"root_causes": [
"Missing implementation of `foreach_trunc` for complex64 tensors on XPU.",
"Lack of support for 'trunc' operation on complex inputs in the current setup."
],
"suggested_solutions": [
"Implement the `foreach_trunc` operation for complex64 tensors on XPU.",
"Add necessary kernel functions to support 'trunc' and related operations for complex numbers."
]
}
Triage bot response: {
"similar_issue_id": 469,
"similar_issue_state": "closed",
"issue_owner": "PenghuiCheng",
"issue_description": "The unit test for 'foreach_trunc' operation on XPU with ComplexDouble type is failing due to the absence of the 'foreach_unary_op_xpu' implementation for complex numbers. The error indicates that 'trunc' is not supported for complex inputs, which is consistent with issues previously reported in Torch-XPU operations.",
"root_causes": [
"The 'foreach_trunc' operation is not implemented for ComplexDouble on XPU.",
"Lack of support for complex number operations in certain XPU unary functions."
],
"suggested_solutions": [
"Implement the 'foreach_trunc' kernel for ComplexDouble on XPU to handle complex numbers.",
"Ensure that all necessary unary operations are supported for complex types on XPU to prevent similar errors in the future."
]
}
Triage bot response: {
"similar_issue_id": 628,
"similar_issue_state": "closed",
"issue_owner": "PenghuiCheng",
"issue_description": "The test case involves creating a tensor using torch.randn with specific batch and vocabulary sizes, then converting it to bfloat16. The code is as follows: import torch; assert torch.xpu.is_available(), 'Intel XPU is not available'; batch_size = 4; vocab_size = 4; out = torch.randn(batch_size, vocab_size, device='xpu').to(torch.bfloat16)\nx = torch.randn(4, 4, dtype=torch.float32, device='xpu')\\nx.to(torch.float16)",
"root_causes": [
"Lack of support for specific data type conversions on XPU devices, particularly involving fp64.",
"Kernel implementation for data conversion operations lacks paths for fp64 casting.",
"Inadequate handling of complex number operations on XPU devices."
],
"suggested_solutions": [
"Modify the torch.randn() implementation to use fp32 as an intermediate type when fp64 is not supported on the device.",
"Implement support for complex number operations in the XPU backend.",
"Ensure that all relevant functions, such as trunc, are properly supported for complex inputs on XPU devices."
]
}
Triage bot response: {
"similar_issue_id": 845,
"similar_issue_state": "closed",
"issue_owner": "daisyden",
"issue_description": "The test third_party.torch-xpu-ops.test.xpu.test_nn_xpu.TestNN/test_LayerNorm_3d_no_affine_large_feature_cuda failed with an AssertionError: Tensor-likes are not close! The error suggests a mismatch in tensor values between CUDA and XPU implementations, potentially due to differences in precision or implementation. The test involves a LayerNorm operation in a 3D setup with a large feature dimension without affine layers.",
"root_causes": [
"Discrepancies in LayerNorm implementation between CUDA and XPU leading to tensor value mismatches.",
"Potential precision handling differences affecting gradient computations."
],
"suggested_solutions": [
"Align XPU LayerNorm implementation with CUDA to ensure consistent results.",
"Review and adjust precision handling to match CUDA's behavior."
]
} |
Signed-off-by: Cheng, Penghui <[email protected]>
|
LGTM |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add UT error cases to skip list for LNL windows device.
report of the issues :
#1733
#1737
#1739