Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[ARC Windows] UT failures #910

Open
22 tasks
mengfei25 opened this issue Sep 13, 2024 · 3 comments
Open
22 tasks

[ARC Windows] UT failures #910

mengfei25 opened this issue Sep 13, 2024 · 3 comments
Assignees
@min-jean-cho
Labels
client ut_triage Windows
Milestone
PT2.5

Comments

@mengfei25
Copy link
Contributor

🐛 Describe the bug

Failures on ARC windows, total 2127,FP64 related issue: 1910, others: 217

  • AssertionError: 'Assertion cur_target >= 0 && cur_target < n_classes failed' not found in 'PYTORCH_API_USAGE
  • AssertionError: "Kernel\ is\ incompatible\ with\ all\ devices\ in\ devs" does not match "Required aspect fp64 is not supported on the device"
  • AssertionError: "not implemented for" does not match "Native API failed. Native API returns: -999 (Unknown PI error) -999 (Unknown PI error)"
  • AssertionError: RuntimeError not raised
  • AssertionError: Scalars are not close!
  • AssertionError: Tensor-likes are not close!
  • AssertionError: Tensor-likes are not equal!
  • Exception: Caused by sample input at index
  • RuntimeError: Caught RuntimeError in DataLoader worker process 1.
  • RuntimeError: Comparing
  • RuntimeError: could not create a primitive descriptor for a convolution forward propagation primitive
  • RuntimeError: could not create a primitive descriptor for a deconvolution forward propagation primitive
  • RuntimeError: Default context is not supported on XPU on Windows. So we can NOT find its global index of the ATen device.
  • RuntimeError: Kernel is incompatible with all devices in devs
  • RuntimeError: Loader error
  • RuntimeError: Native API failed. Native API returns: -999 (Unknown PI error) -999 (Unknown PI error)
  • RuntimeError: Ninja is required to load C++ extensions
  • RuntimeError: Required aspect fp64 is not supported on the device
  • RuntimeError: Worker error
  • RuntimeError: XPU out of memory, please use empty_cache to release all unoccupied cached memory.
  • AssertionError: Scalars are not close!
  • Exception: Caused by sample input at index

Versions

torch-xpu-ops: release/2.5
@chuanqi129 chuanqi129 added this to the PT2.5 milestone Sep 18, 2024
@chuanqi129
Copy link
Contributor

@min-jean-cho could you please double check whether those issues are expected?

@chuanqi129
Copy link
Contributor

@PenghuiCheng please share the linux ARC UT test results as reference

@PenghuiCheng
Copy link
Contributor

@PenghuiCheng please share the linux ARC UT test results as reference

attached. arc_log.txt is UT of test/xpu, and arc_extended_log.txt is UT of test/xpu/extended.
arc_log.txt
arc_extended_log.txt

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@min-jean-cho min-jean-cho

Labels
client ut_triage Windows
Projects
None yet
Milestone
PT2.5
Development

No branches or pull requests
4 participants
@chuanqi129 @PenghuiCheng @mengfei25 @min-jean-cho