[CI] Run extended test dispatch on internal runner

[gyohuangxin]

Update the Extended Test dispatch job to run on the internal linux-aiter-build-mi250 runner instead of GitHub hosted ubuntu-latest. This lets the dispatch call reach the private AMD-ROCm-Internal/fw-bringup repository from an allowed network path.

[Copilot]

Copilot wasn't able to review any files in this pull request.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[github-actions]

🏷️ CI Guide

Runs automatically on every PR:

• ✅ Pre-checks (submodule verification, code formatting)
• ✅ Aiter op tests (gfx942 + gfx950)
• ✅ Triton tests on MI35X (only when aiter/ops/triton/** or related paths are changed)

Extended tests (opt-in via labels):

Label	Tests
ci:gfx1250-ffm-triton	Run the five-shard gfx1250 FFM Triton test suite
ci:triton-300x	Run an additional Triton test job on MI300X in PRs; main branch always runs both MI35X and MI300X
multigpu	Aiter multi-GPU tests on the 8-GPU runner
ci:sglang	SGLang integration tests: DeepSeek-R1-MXFP4 accuracy, Qwen 3.5 accuracy
ci:atom	ATOM benchmark: DeepSeek-R1-0528, GPT-OSS-120B
ci:atom_full	ATOM accuracy suite for PR and main models from ATOM models_accuracy.json
ci:vllm	vLLM benchmark: GPT-OSS-120B, DeepSeek-R1-0528, Kimi-K2.5
ci:all	All standard extended tests (excludes ci:atom_full)

Only add ci:atom_full for FlyDSL or Triton upgrades.
Add labels via the sidebar or gh pr edit 5205 --add-label <label>

PR title tags & labels:
Component tags ([Triton/Gluon], [HIP], [CK], [ASM], ...) are added to the PR title and as PR labels automatically from the changed files and re-synced on every push — change-type tags like [fix]/[Perf], op tags like [MLA], and human labels (ci:*) are left untouched. Add the no-auto-title label to opt this PR out.

[Copilot]

🟡 Changes recommended

The workflow now runs a dispatch-only job on an MI250 self-hosted runner, which can unnecessarily consume GPU runner capacity and impact CI throughput.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Review details

• Files reviewed: 1/1 changed files
• Comments generated: 1
• Review effort level: Lite

[Copilot]

🔵 Needs a closer look

Moving the long-polling “wait for status” loop onto the internal MI250 runner can unnecessarily tie up scarce self-hosted capacity for up to ~180 minutes per run.

Review details

Suppressed comments (1)

.github/workflows/extended-test.yaml:33

• Switching this job to the internal linux-aiter-build-mi250 runner will now keep that self-hosted runner occupied for up to EXTENDED_TEST_WAIT_TIMEOUT_MINUTES (default 180) due to the polling loop in “Wait for extended test result” (sleeping/retrying every 30s). Since this job mostly performs GitHub API calls, consider splitting into two jobs: (1) a short dispatch job on the internal runner, and (2) a separate wait/poll job on a GitHub-hosted runner that needs: the dispatch job. This avoids tying up scarce MI250 capacity for hours waiting on external status.

        contains(github.event.pull_request.labels.*.name, 'ci:extended-test')
      ) &&
      (github.event.action != 'labeled' || github.event.label.name == 'ci:extended-test')
    runs-on: linux-aiter-build-mi250

• Files reviewed: 2/2 changed files
• Comments generated: 0 new
• Review effort level: Lite

[leo-automation]

LGTM