[xpu][test] Skip WIP config for Intel GPU in test_safetensors_support.py and test_x86inductor_fusion.py by zxd1997066 · Pull Request #4049 · pytorch/ao

zxd1997066 · 2026-03-11T08:11:54Z

For #2917, This PR is targeted to skip WIP config for Intel GPU in test/prototype/safetensors/test_safetensors_support.py and test/quantization/pt2e/test_x86inductor_fusion.py

pytorch-bot · 2026-03-11T08:11:57Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4049

📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (1 Unrelated Failure)

As of commit fd653f3 with merge base fe98658 ():

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

Run Regression Tests / test-nightly (CUDA Nightly, linux.g5.12xlarge.nvidia.gpu, --pre torch --index-url https://downloa... / linux-job (gh) (trunk failure)
test/test_low_bit_optim.py::TestFSDP2::test_fsdp2

This comment was automatically generated by Dr. CI and updates every 15 minutes.

pytorch-bot · 2026-03-19T01:57:55Z

To add the ciflow label ciflow/xpu please first approve the workflows that are awaiting approval (scroll to the bottom of this page).

This helps ensure we don't trigger CI on this PR until it is actually authorized to do so. Please ping one of the reviewers if you do not have access to approve and run workflows.

xiaowangintel · 2026-03-19T01:58:53Z

@liangan1 Please help to review this pr. It block xpu-ci.

xiaowangintel · 2026-03-19T02:22:50Z

@Stonepia Please help to review.

liangan1 · 2026-03-19T05:23:52Z

@@ -63,7 +63,7 @@
    ]



XPU uses a different path for INT4 woq in the line 60., so we add skip here.

@liangan1 It has been enabled https://github.com/zxd1997066/ao/blob/3fde9ea2707775d23b773a5bb4e8afeaffc00aa6/test/prototype/safetensors/test_safetensors_support.py#L60

Yes. you are right. We need to test XPU on https://github.com/zxd1997066/ao/blob/3fde9ea2707775d23b773a5bb4e8afeaffc00aa6/test/prototype/safetensors/test_safetensors_support.py#L60 and skip line 52 which is only for CUDA.

liangan1 · 2026-03-19T05:24:17Z


 # MX and NVFP4 configs require torch >= 2.11
-if torch_version_at_least("2.11.0.dev"):
+if not torch.xpu.is_available() and torch_version_at_least("2.11.0.dev"):


MX format is not enable yet on Intel XPU.

jerryzh168 · 2026-03-24T00:41:44Z

 # Int4WeightOnlyConfig with tinygemm uses sm90a CUTLASS kernels which are not
 # forward-compatible with sm100+
-if not is_sm_at_least_100():
+if not torch.xpu.is_available() and not is_sm_at_least_100():


these xpu conditions are spread across the code base and makes it confusing on what is and is not supported on cuda and xpu, can you just have 1 branch for xpu and one branch for not xpu to make things clearer

Thanks, refined the code.

jerryzh168 · 2026-03-24T02:56:27Z

also this is pretty weird, why do we get the last device in _DEVICES?

removed this line.

jerryzh168 · 2026-03-24T02:57:40Z

I think you'll want to refactor this to use get_available_devices as well, like other files

Thanks, updated

jerryzh168 · 2026-03-24T02:58:56Z

-    ]
-
 # plain_int32 only supports XPU
 if _DEVICE in ("xpu"):


probably should be if torch.xpu.is_available() to be consistent with other conditions

Thanks, updated

jerryzh168 · 2026-03-24T03:01:14Z

    @unittest.skipIf(
        torch_version_at_least("2.11.0.dev"), "Doesn't work with torch 2.11.0.dev+"
    )
+    @unittest.skipIf(torch.xpu.is_available(), "Doesn't support for XPU")


nit: Doesn't work with XPU

meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Mar 11, 2026

zxd1997066 mentioned this pull request Mar 11, 2026

[TorchAO][UT] failures in test/prototype/safetensors/test_safetensors_support.py intel/torch-xpu-ops#3032

Closed

xiaowangintel added ciflow/xpu label used to trigger xpu CI jobs xpu Intel XPU related features module: core changes affecting multiple modules, e.g. base config/tensor, observers, quant ops labels Mar 19, 2026

pytorch-bot Bot removed the ciflow/xpu label used to trigger xpu CI jobs label Mar 19, 2026

xiaowangintel requested a review from liangan1 March 19, 2026 01:58

xiaowangintel approved these changes Mar 19, 2026

View reviewed changes

xiaowangintel added the ciflow/xpu label used to trigger xpu CI jobs label Mar 19, 2026

Stonepia approved these changes Mar 19, 2026

View reviewed changes

liangan1 approved these changes Mar 19, 2026

View reviewed changes

liangan1 requested a review from jerryzh168 March 19, 2026 05:24

zxd1997066 force-pushed the xiangdong/skip_safetensors branch from 3fde9ea to bc7893b Compare March 19, 2026 05:36

pytorch-bot Bot removed the ciflow/xpu label used to trigger xpu CI jobs label Mar 19, 2026

zxd1997066 changed the title ~~[xpu][test] Skip WIP config for Intel GPU in test/prototype/safetensors/test_safetensors_support.py~~ [xpu][test] Skip WIP config for Intel GPU in test_safetensors_support.py and test_x86inductor_fusion.py Mar 19, 2026

zxd1997066 force-pushed the xiangdong/skip_safetensors branch from bc7893b to 01d473c Compare March 19, 2026 05:41

jerryzh168 reviewed Mar 24, 2026

View reviewed changes

zxd1997066 added 3 commits March 24, 2026 10:38

skip WIP config for XPU

2a7c770

skip cases for xpu in test/quantization/pt2e/test_x86inductor_fusion.py

f2f7707

refine the code

dbf1641

zxd1997066 force-pushed the xiangdong/skip_safetensors branch from 01d473c to dbf1641 Compare March 24, 2026 02:38

liangan1 requested review from liangan1 and xiaowangintel March 24, 2026 02:45

liangan1 approved these changes Mar 24, 2026

View reviewed changes

liangan1 added the ciflow/xpu label used to trigger xpu CI jobs label Mar 24, 2026

liangan1 requested a review from jerryzh168 March 24, 2026 02:46

jerryzh168 reviewed Mar 24, 2026

View reviewed changes

refine the code

5c1a40c

pytorch-bot Bot removed the ciflow/xpu label used to trigger xpu CI jobs label Mar 24, 2026

refine skip reason

fd653f3

zxd1997066 requested a review from jerryzh168 March 24, 2026 05:25

jerryzh168 approved these changes Mar 24, 2026

View reviewed changes

jerryzh168 added the ciflow/xpu label used to trigger xpu CI jobs label Mar 24, 2026

liangan1 mentioned this pull request Mar 25, 2026

[reland][xpu] INT8 quantization on Intel XPU #3782

Merged

liangan1 merged commit 8fca033 into pytorch:main Mar 25, 2026
20 of 21 checks passed

Freed-Wu mentioned this pull request Apr 12, 2026

Add torch.uint16, torch.uint32 #4269

Open

Conversation

zxd1997066 commented Mar 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot Bot commented Mar 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4049

✅ You can merge normally! (1 Unrelated Failure)

Uh oh!

pytorch-bot Bot commented Mar 19, 2026

Uh oh!

xiaowangintel commented Mar 19, 2026

Uh oh!

xiaowangintel commented Mar 19, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jerryzh168 Mar 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

zxd1997066 commented Mar 11, 2026 •

edited

Loading

pytorch-bot Bot commented Mar 11, 2026 •

edited

Loading

jerryzh168 Mar 24, 2026 •

edited

Loading