List view

[distributed] AttributeError: module 'torch.xpu' has no attribute '_sleep'

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

Feature

#1727 In intel/torch-xpu-ops;

· PenghuiCheng opened on Jun 9, 2025

[distributed] ValueError: Cannot use ReduceOp.PREMUL_SUM with XCCL

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

Feature

#1571 In intel/torch-xpu-ops;

· daisyden opened on Apr 11, 2025

[Evaluated] FP8 support in matmul // Issues in test_matmul_cuda.py due to FP8

triaged

Status: Open.

Feature

#322 In intel/torch-xpu-ops;

· yuchengliu1 opened on May 25, 2024

implement torch.linalg.cholesky xpu backend

bug_fix_stage2

module: op impl

Status: Open.

Feature

#1936 In intel/torch-xpu-ops;

· jiqing-feng opened on Aug 15, 2025

implement torch.linalg.pinv xpu backend

bug_fix_stage2

module: op impl

Status: Open.

Feature

#1902 In intel/torch-xpu-ops;

· yao-matrix opened on Aug 1, 2025

[upstream_ut] segfault with test_fake_crossref_backward_no_amp_to_sparse_xpu_float32

dependency component: driver

module: ut

skippedUsed for temp UT failure to parallel fixUsed for temp UT failure to parallel fix

Status: Open.

#1962 In intel/torch-xpu-ops;

· daisyden opened on Aug 26, 2025

Accuracy issue in test_transformer.py cases

module: op impl

triaged

Status: Open.

Feature

#761 In intel/torch-xpu-ops;

· PenghuiCheng opened on Aug 14, 2024

implement torch.linalg.qr xpu backend

bug_fix_stage2

module: op impl

Status: Open.

Feature

#1900 In intel/torch-xpu-ops;

· yao-matrix opened on Aug 1, 2025

result of bert positional embed changed in 2.8

Accuracy

bug_fix_stage2

Status: Open.

Bug

#1811 In intel/torch-xpu-ops;

· jianyizh opened on Jul 4, 2025

inf is returned by nn.TransformerEncoderLayer

skippedUsed for temp UT failure to parallel fixUsed for temp UT failure to parallel fix

Status: Open.

#2015 In intel/torch-xpu-ops;

· daisyden opened on Sep 4, 2025

[distributed][shared_tensor] test\distributed\_shard\shared_tensor\test_sharded_tensor.py has 12 cases failed with "RuntimeError: eof (this error originated at tensorpipe/transport/shm/connection_impl.cc:259)"

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#2004 In intel/torch-xpu-ops;

· libohao1201 opened on Sep 2, 2025

[distributed]RuntimeError: Work ran time out after 15370 milliseconds

module: distributedFor distributed feature issueFor distributed feature issue

regression

Status: Open.

Bug

#1704 In intel/torch-xpu-ops;

· PenghuiCheng opened on May 27, 2025

[distributed] RuntimeError: Work ran time out after 18928 milliseconds.

bugSomething isn't workingSomething isn't working

duplicateThis issue or pull request already existsThis issue or pull request already exists

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

Bug

#1728 In intel/torch-xpu-ops;

· PenghuiCheng opened on Jun 9, 2025

[distributed] NotImplementedError: The operator 'aten::_scaled_mm.out' is not currently implemented for the XPU device.

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#1550 In intel/torch-xpu-ops;

· PenghuiCheng opened on Apr 7, 2025

[distributed] Some test_fsdp_core.py cases got random failures in _join_processes(fn)

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#1475 In intel/torch-xpu-ops;

· daisyden opened on Mar 17, 2025

[distributed][_composable/test_composability] RuntimeError: 'Attempted to flatten sharded dimension 1

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#1881 In intel/torch-xpu-ops;

· zxd1997066 opened on Jul 25, 2025

segmentation while lora finetuning

dependency component: driver

dependency component: oneDNN

hw: PVC

Status: Open.

Bug

#1810 In intel/torch-xpu-ops;

· yao-matrix opened on Jul 4, 2025

New huggingface LLM models issues

bug_fix_stage2

E2E

hw: PVC

Status: Open.

#2055 In intel/torch-xpu-ops;

· mengfei25 opened on Sep 17, 2025

reduction got accuracy issue on large tensor cases

module: ut

regression

skippedUsed for temp UT failure to parallel fixUsed for temp UT failure to parallel fix

Status: Open.

Bug

#2008 In intel/torch-xpu-ops;

· daisyden opened on Sep 2, 2025

[distributed][elastic/rendezvous] FileNotFoundError: [Errno 2] No such file or directory: 'etcd'

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#2124 In intel/torch-xpu-ops;

· zxd1997066 opened on Sep 29, 2025

[distributed][fsdp]Accuracy gaps on xelink

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#1926 In intel/torch-xpu-ops;

· zxd1997066 opened on Aug 12, 2025

[distributed] test_device_mesh.py::TestDeviceMeshGetItem::test_flatten_mesh_3d AssertionError

bugSomething isn't workingSomething isn't working

module: distributedFor distributed feature issueFor distributed feature issue

Status: Open.

#2165 In intel/torch-xpu-ops;

· zxd1997066 opened on Oct 14, 2025

Reduce the character number of linkage command

windows_ciOnly for Windows CI triggerOnly for Windows CI trigger

Status: Open (in progress).

intel/torch-xpu-opsnumber 1303

#1303 In intel/torch-xpu-ops;

· CuiYifeng opened on Jan 20, 2025

convit_base bfloat16 training accuracy got eager_fail_to_run randomly

Accuracy

bug_fix_stage3

E2E

Status: Open.

#2150 In intel/torch-xpu-ops;

· mengfei25 opened on Oct 10, 2025

BMG d2h copy is very slow compare to pvc and 4080s

dependency component: driver

performance

Status: Open.

#2157 In intel/torch-xpu-ops;

· jianyizh opened on Oct 10, 2025