Skip to content

[XPU][Rebase v0.22] Fix for 0.22 rebase #4059

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

Merged
hsliuustc0106 merged 1 commit into from
Jun 2, 2026
+9 −4
Merged

[XPU][Rebase v0.22] Fix for 0.22 rebase #4059

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .buildkite/pipeline-intel.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@ steps:
DOCKER_BUILDKIT: "1"
# Buildkite will automatically replace this with the actual commit hash
VLLM_IMAGE_TAG: "${BUILDKITE_COMMIT}"
VLLM_VERSION: "v0.21.0"
VLLM_VERSION: "v0.22.0"
priority: 100
timeout_in_minutes: 60
soft_fail: false
3 changes: 2 additions & 1 deletion tests/e2e/offline_inference/test_qwen2_5_omni_expansion.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,7 +39,8 @@ def get_cuda_graph_config():
stage_config = get_cuda_graph_config()

# Create parameter combinations for model and stage config
test_params = [(model, stage_config) for model in models]
# Qwen2.5-Omni with TP=3 needs longer init timeout
test_params = [(model, stage_config, {"stage_init_timeout": 1200, "init_timeout": 1800}) for model in models]


def get_question(prompt_type="mix"):
Expand Down
8 changes: 6 additions & 2 deletions tests/helpers/mark.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,10 @@ def cuda_marks(*, res: str, num_cards: int):
return marks
test_distributed = pytest.mark.distributed_cuda(num_cards=num_cards)

if not current_platform.is_cuda():
return marks + [test_distributed]
Comment on lines +24 to +25

@chatgpt-codex-connector chatgpt-codex-connector Bot Jun 1, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Preserve a skip when no requested platform is present

When this helper is evaluated on a non-CUDA host it now returns only marker tags, with no skipif. For multi-platform tests where the other platform helpers also do not add a skip on the current host (for example test_qwen2_5_omni_expansion.py on a CPU-only or otherwise unsupported runner), the test is collected with no active skip and will try to start the large hardware-only model instead of being skipped. The intended XPU fix needs a cross-platform skip only when the current platform is not any of the resources requested, while still avoiding CUDA's skip from suppressing XPU runs.

Useful? React with 👍 / 👎.

test_skipif = pytest.mark.skipif(
not current_platform.is_cuda() or (current_platform.device_count() < num_cards),
current_platform.device_count() < num_cards,
reason=f"Need at least {num_cards} CUDA GPUs to run the test.",
)
return marks + [test_distributed, test_skipif]
Expand Down Expand Up @@ -52,8 +54,10 @@ def xpu_marks(*, res: str, num_cards: int):
return marks
test_distributed = pytest.mark.distributed_xpu(num_cards=num_cards)

if not current_platform.is_xpu():
return marks + [test_distributed]
test_skipif = pytest.mark.skipif(
not current_platform.is_xpu() or (current_platform.device_count() < num_cards),
current_platform.device_count() < num_cards,
reason=f"Need at least {num_cards} XPUs to run the test.",
)
return marks + [test_distributed, test_skipif]
Expand Down
Loading