Skip to content

chore(xpu): bump vllm-openai-xpu to v0.33.0 - #96

Open
aung-san-i wants to merge 4 commits into
mainfrom
deps/xpu-vllm-v0.33.0
Open

aung-san-i wants to merge 4 commits into
mainfrom
deps/xpu-vllm-v0.33.0

Conversation

@aung-san-i

Copy link
Copy Markdown
Owner

Adopts vllm/vllm-openai-xpu:v0.33.0 as the XPU base and runtime image.

Image digest sha256:be3a9aea234be64ffc2816ac561c597e0f0b1f44a61a7f4912780e4d77450086
New licence baseline vllm-openai-xpu@be3a9aea
Platform linux/amd64
Capture run https://github.com/aung-san-i/dynamo/actions/runs/41490948555

What changed

  • container/context.yaml — base_image_tag, runtime_image_tag
    and baseline_sbom in the vllm: xpu: block. Three lines, verified
    as exactly three before this PR was opened.
  • container/compliance/base_sboms/ — the inventory for the new image,
    with the superseded one deleted.

Why the baseline moves too

The licences stage subtracts a stored inventory of the base image, and
the subtraction matches on (name, version). Bumping the tag alone leaves
every package whose version moved unsubtracted, which reads as net-new
and trips the policy gate at container/templates/compliance.Dockerfile
line 125 — so build-xpu would fail on a licence error and no test
would run. Re-capturing in the same commit is what keeps the gate honest.

What to check before merging

  1. xpu-status-check is green — that is the real signal that this
    release works on XPU.
  2. The container/context.yaml diff touches only the xpu: block.
  3. The baseline re-capture is a compliance statement about what we
    treat as upstream's responsibility, so it wants a human read.

Not auto-merged on purpose. Opened by
.github/workflows/xpu-base-image-watch.yaml.

dynamo-ops and others added 2 commits November 2, 2026 17:36
Re-captures the amd64 licence baseline in the same commit as the tag
bump. The licences stage subtracts on (name, version), so bumping the
tag alone leaves every moved package unsubtracted and fails the policy
gate in compliance.Dockerfile before any test can run.

Image digest: sha256:be3a9aea234be64ffc2816ac561c597e0f0b1f44a61a7f4912780e4d77450086
Baseline stem: vllm-openai-xpu@be3a9aea

Opened by .github/workflows/xpu-base-image-watch.yaml

Signed-off-by: dynamo-ops <170655669+dynamo-ops@users.noreply.github.com>
…LLM v0.33.0

vLLM v0.33.0's NIXL connector refuses NIXL < 1.5.0, and the XPU image
inherited the framework-level nixl_ref v1.4.1, so every disaggregated
XPU worker exited at engine start (run 41492764713: px-dn, pn-dx,
px-dn-hnd-tp2, px-dx).

vLLM-Omni 0.31.0rc1 imports get_engine_zmq_addresses, which vLLM 0.33
renamed, so the omni DiT worker failed on import (omni-ar-dit-2nodes).
No Omni release targets vLLM 0.33 yet; 0.32.0 has the renamed import.

Transformers 5.14.1 / tokenizers 0.22.2 are below vLLM 0.33's floors
(>=5.16.1, >=0.23.2). 5.19.0 is the newest release inside both vLLM
0.33's range (<5.21) and Omni 0.32's (<5.20).

All three pins are scoped to vllm.xpu; CUDA and CPU are unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aung-san-i and others added 2 commits November 2, 2026 14:48
Signed-off-by: Aung San aka Mike <41126828+aung-san-i@users.noreply.github.com>
The merge of main dropped vllm.xpu.nixl_ucx_ref, so the XPU image fell
back to the dynamo-wide nixl_ucx_ref v1.22.0. main builds XPU with UCX
v1.23.0-rc1 next to NIXL 1.5.0 (args.Dockerfile reads the per-device
value first), so keep that pairing here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants