docs(gateway): describe tensor-parallel device groups in the architecture guide

[krisztian-gajdar]

What this changes

The gateway architecture guide's "Multi-GPU caveat" still said a generation model is never sharded across GPUs. Since #282, a pool with gpu.deviceGroup: true renders one worker child that owns every GPU in the pod, and a profile that declares tensor_parallel_size runs one engine across those GPUs.

The paragraph now says what is true:

• Routing still binds a model to one child and does not replicate it across children.
• Sharding happens inside that child.

Docs only. The wording matches the "Tensor-Parallel Device Groups" section of deploy/helm/sie-cluster/README.md.

Summary by CodeRabbit

• Documentation
  • Clarified multi-GPU generation behavior: models may shard across multiple GPUs within a child when GPU device grouping and tensor parallelism are configured.
  • Clarified that the gateway and sidecar continue to treat the child as a single worker.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Team

Run ID: 5f160456-4dff-4c58-9152-52e702ef9ec1

📥 Commits

Reviewing files that changed from the base of the PR and between 7a8b09b and 50e2025.

📒 Files selected for processing (1)

• packages/sie_gateway/docs/architecture-guide.md

Included review availability: Your plan provides up to 10 included reviews per hour; 7 remain after this review.

---

📝 Walkthrough

Walkthrough

The architecture guide now states that a model remains assigned to one child but may use multiple GPUs within that child when device grouping and tensor parallelism are configured.

Changes

Generation placement

Layer / File(s)	Summary
Clarify intra-child model placement packages/sie_gateway/docs/architecture-guide.md	The guide now distinguishes child-level assignment from intra-child tensor parallelism. It removes the statement that one model cannot shard across multiple GPUs.

Suggested reviewers: huronat

Priority: ⬇️ Low

Merge Risk: ⚪ Minimal · up to 50e20

This documentation clarification introduces no established production or deployment risk.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the documentation change about tensor-parallel device groups in the gateway architecture guide.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

---

_{Comment @coderabbitai help to get the list of available commands.}