New failed test cases 2026-04-02

[Silv3S]

🐛 Describe the bug

Cases:
op_ut,third_party.torch-xpu-ops.test.xpu.nn.test_convolution_xpu.TestConvolutionNNDeviceTypeXPU,test_conv2d_hipdnn_backend_selection_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.nn.test_convolution_xpu.TestConvolutionNNDeviceTypeXPU,test_conv2d_hipdnn_deterministic_error_xpu

Versions

2.12

[github-actions]

✅ op_ut,third_party.torch-xpu-ops.test.xpu.test_transformers_xpu.TestSDPAFailureModesXPU,test_mem_eff_attention_large_seq_len_uniform_attention_xpu Passed in nightly testing

[kaileiyx]

Cases:
op_ut,third_party.torch-xpu-ops.test.xpu.test_transformers_xpu.TestSDPAFailureModesXPU,test_mem_eff_attention_large_seq_len_uniform_attention_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_transformers_xpu.TestSDPAFailureModesXPU,test_mem_eff_attention_large_seq_len_uniform_attention_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_transformers_xpu.TestSDPAFailureModesXPU,test_mem_eff_attention_large_seq_len_uniform_attention_xpu

test_mem_eff_attention_large_seq_len_uniform_attention_xpu failed in the latest nightly run. It's a regression
https://github.com/intel/torch-xpu-ops/actions/runs/24346789650/job/71171973837

error:

[gw8] [ 19%] RERUN test_transformers_xpu.py::TestSDPAFailureModesXPU::test_mem_eff_attention_large_seq_len_uniform_attention_xpu 

replacing crashed worker gw8

[Silv3S]

Moving these 3 test cases to separate ticket and marking as random #3326